| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2003;126:969-975
© 2003 The American Association for Thoracic Surgery
Surgery for acquired cardiovascular disease |
a Division of Cardiac Surgery, Department of Surgery, Brigham & Women's Hospital, Harvard Medical School, Boston, Mass, USA
Read at the Eighty-third Annual Meeting of The American Association for Thoracic Surgery, Boston, Mass, May 4-7, 2003.
Received for publication March 13, 2003; revisions received April 14, 2003; accepted for publication April 21, 2003.
* Address for reprints: Lawrence H. Cohn, MD, Division of Cardiac Surgery, Brigham & Women's Hospital, 75 Francis St, Boston, MA 02115, USA
lcohn{at}partners.org
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
METHODS: We performed a retrospective cohort study on 144 patients at a single institution who had bioprosthetic aortic or mitral valves removed, serum cholesterol levels recorded, and valve calcification assessed on the basis of hematoxylin and eosin staining and radiography of the valve. We also performed case-control analysis of a group of 66 patients whose tissue valves were explanted and compared them with an age- and position-matched group of 66 patients with similar duration of implantation. We also compared mean serum cholesterol levels.
RESULTS: In the retrospective cohort study cholesterol (P = .035), younger age at implantation (P = .014), and coronary artery disease (P = .017) were linked to calcification of the valve by means of univariate analysis. In stepwise multiple regression analysis only the mean serum cholesterol level was linked to calcification (P = .02). Sex, hypertension, smoking, diabetes, and implant position were not linked to calcification. In the case-control analysis the mean serum cholesterol level of the explanted valve group was significantly higher (189 vs 163 mg/dL, P < .0001) than that of the group whose valves did not require explantation. For those whose serum cholesterol levels were greater than 200 mg/dL, the odds ratio was 3.9 (95% confidence interval, 1.7-8.9) for valve explantation.
CONCLUSIONS: Increased serum cholesterol level may be a risk factor for bioprosthetic valve calcification requiring explantation.
|
Bioprosthetic valve replacement has improved the outcomes of patients with valvular heart disease. Tissue valves have central flow and a high freedom from thrombus formation, usually without the need for anticoagulation. However, most tissue valves are altered biologic materials that undergo degenerative processes.1,2
A study of the causes of valve failure necessitating reoperation at our institution has reported that sterile degeneration, or primary tissue failure, accounted for 74% of bioprosthetic valve failures.3 Primary tissue failure was most often related to calcification of the tissue cusps. Calcification was present in 86% of patients, and cuspal defects in 79% of bioprosthetic valves were removed for primary tissue failure. Seventy-five percent of valves with degenerative failure had both calcification and tears.
No medical therapy, as of yet, has been demonstrated to be effective in the prevention of the progression of bioprosthetic calcification. Recently, a compelling argument has been espoused demonstrating the similarities between calcific aortic sclerosis in the native valve and atherosclerosis.4,5 Chronic inflammatory changes in the leaflet, including cellular infiltration of macrophages and T lymphocytes and lipoprotein deposition in apposition to areas of calcification, has prompted investigation as to whether medical therapy can be applied to slow the progression of valvular failure.
Several studies have documented a correlation between lipids and calcification. Pohle and colleagues6 have documented that both coronary calcification and aortic valve calcification progress more rapidly in subjects with low-density lipoprotein (LDL) levels of greater than 130 mg/dL. Callister and associates7 demonstrated that patients who reduced their cholesterol levels with lipid-lowering agents significantly reduced the progression of coronary calcification. In another study the risk factor that most significantly correlated with coronary calcification was LDL levels, and this was stronger than age, male sex, blood pressure, fasting insulin level, or smoking.8 Native aortic valve calcium accumulation, as assessed by means of electron beam computed tomography, is diminished by statin use.9
We hypothesized that serum cholesterol levels might be associated with increasing bioprosthetic valve calcification, as assessed by means of a retrospective cohort study. We also questioned whether increased serum cholesterol levels are associated with a higher incidence of valve explantation, as assessed by means of a case-control study.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Historical cohort
We performed a retrospective cohort study of all patients referred to our single center over 10 years (1992-2002) for individuals who had their bioprosthetic aortic or mitral valves removed and who had serum cholesterol levels (excluding up to 1 month after valve surgery) measured during the period when the valve was in use. One hundred forty-four patients met the inclusion and exclusion criteria. We reviewed the medical records for the pathologic assessment of calcification at the time of explantation. Calcification was assessed by means of a combination of histologic and radiographic methods. Valves were radiographed for calcium and stained with hematoxylin and eosin after decalcification of the tissue for sectioning. A grading scheme of 0 to 4+ was reported, with heavy calcification given a rating of 4+ and no calcification given a rating of 0.
Case-control study
We performed a case-control study of patients referred to our single center over 10 years (1992-2002) for individuals who had their bioprosthetic aortic or mitral valves removed and who had serum cholesterol levels measured while the valve was in place. Control subjects were matched to patients for (1) duration of valve survival, (2) age of the patient at the time of implantation (within 1-5 years of patient age), and (3) valve position (aortic vs mitral) by using the Brigham Cardiac Surgery Database. We matched patients against our computerized prospective database from 1992 onward, using a computerized sorting algorithm that first sorted by duration of survival (for control subjects) or removal (for patients), then by age at implantation, and finally by position. Fewer patients are enrolled in the case-control analysis than in the cohort analysis because of adequate statistical power with approximately 70 patients in each group and difficulty in finding adequate control subjects for all patients. It should be noted that the same patients in the cohort analysis were used as patients here. Matched control subjects, however, are a different group of patients and are not represented in the group of patients in the cohort analysis.
Patients and control subjects were temporally mismatched, on average, by 6 years because the mean ± SD year of birth for the patients was 1930 ± 1.8 versus 1936 ± 1.7 for control subjects (P < .001). Thus the age of the patient at implantation is matched, but patients represent an earlier surgical population. We performed 2 additional analyses to investigate whether this had any effect on mean cholesterol levels between patients and control subjects. First, we removed all pairs in which one or both members were taking statins to remove the confounding effects of these drugs (data not shown). Second, we reanalyzed patients by matching for year of operation (thus valves were implanted during the same year in both patients and control subjects), age of patient at implantation, and position. This effectively removed any possible confounding effects of changing medical-surgical practices, although it reduced our analyses to 22 patients.
Statistical analyses
All analyses were carried out with the Statistica software package (version 5.5, Tulsa, Okla), and P values are reported. For cohort analyses, we used the nonparametric test of correlation because calcification was not distributed in a normal manner (Table 1 for the distribution of calcification). We used the Spearman rank order coefficient to compare ordered categoric data (eg, calcification) versus continuous data (eg, serum cholesterol level). To compare ordered categorical versus dichotomous data (eg, sex), we used the Wilcoxon rank sum test. All risk factors that had a P value of less than .10 in the univariate analyses were also analyzed by means of stepwise multiple regression.
|
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Stepwise multiple regression was performed on all dependent variables with a significance of less than .10 determined by means of univariate analysis. Thus implant age, coronary artery disease, average serum cholesterol level, and diabetes mellitus were used as dependent variables in predicting calcification. Of the 4 dependent variables, only average cholesterol level achieved significance (P = .02) in the multiple regression (Table 2).
|
|
|
|
We also matched patients for the same year of implantation surgery, age at implantation, and valve position (aortic vs mitral). This effectively removed any influence of possible changing medical-surgical practices through time, although it limited our analysis to 22 patients (Table 4). There was still a significant difference in the mean serum cholesterol levels between patients and control subjects (187 mg/dL in patients and 155 mg/dL in control subjects; P = .047, paired t test). Diabetes (P = .059), hypertension (P = .67), smoking (P = .37), coronary artery disease (P = .34), and statin use (P = .96) were not significantly different between the 2 subgroups.
For the complete case-control analysis, the odds ratio for valvular re-replacement in patients whose serum cholesterol levels were greater than 200 mg/dL was 3.9x (95% confidence interval, 1.7-8.9; Table 3).
Discussion
We have demonstrated for the first time that increasing serum cholesterol levels are associated with increased bioprosthetic valvular calcification. Furthermore, patients who had their valves explanted had higher serum cholesterol levels than an age-, duration of survival-, and position-matched group with unexplanted valves (189 vs 163 mg/dL, P < .0001). Finally, the odds ratio for valve explantation and re-replacement was 3.9-fold higher in patients whose serum cholesterol levels were greater than 200 mg/dL.
It is controversial whether increased serum lipid levels correlate with the progression of peripheral atherosclerosis, coronary calcification, and native aortic valve sclerosis.4,10,11 Pohle and colleagues6 recently demonstrated that the rate of change of both coronary and native aortic valve calcification progresses more rapidly in individuals whose serum LDL levels are greater than 130 mg/dL. Palta and coworkers12 also demonstrated that those with a serum cholesterol level of greater than 200 mg/dL had a rate of native aortic valve area reduction roughly twice that of those with a lower cholesterol level (P = .04).
In this cohort study we demonstrate that blood pressure, sex, and smoking are not linked to tissue valve calcification. We also demonstrate that coronary calcification is linked to tissue valve calcification; thus what is harmful to the coronary arteries might be harmful for tissue valves, and what is protective of the coronary arteries might be protective of the valves. Prior reports have concluded that tissue valves are appropriate for patients undergoing mitral valve replacement with coronary artery disease who are more than 60 years of age and for aortic valve replacement in patients who are more than 70 years of age, irrespective of coronary artery disease.13 Patients with subcritical coronary disease who undergo tissue valve replacement might be candidates for statin therapy.
Younger age has previously been demonstrated as a risk factor for leaflet failure.1,14-16 In our study younger age was linked to increasing calcification. Interestingly, the absolute magnitude of the linear association (
= 0.205, P = .014) was similar for younger patients as for cholesterol ( = 0.176, P = .035), implying that the contribution of cholesterol to calcification of the valve is of a similar importance as younger age. In our study diabetes, a well-documented risk factor for atherosclerosis, demonstrated a trend (P = .069) toward significance in univariate analysis. Patients were considered diabetic if they were being actively medically treated, and thus our study might have missed some patients with untreated diabetes. Also, our limited numbers might not have had adequate power to elicit the contributions of these other risk factors for atherosclerosis on valvular calcification.
Bellamy and associates,17 however, recently reported that there was no correlation between the rate of progression of aortic stenosis with cholesterol levels in 157 patients with aortic stenosis. Statin therapy, however, was associated with a decreased rate of progression of stenosis-sclerosis of the aortic valve. In that study the authors confirmed that aortic stenosis progresses with a reduction in orifice size of approximately 0.1 cm2/y in patients who are not receiving statin therapy. For those receiving medical therapy, the decrease in mean aortic valve area measured was only 0.04 cm2/y, which is approximately half the native rate. Palta and coworkers12 also documented a near doubling of the rate of aortic valve area reduction in those whose serum cholesterol levels were greater than 200 mg/dL. After adjustment for sex, age, and baseline aortic valve area, statin therapy remained protective, although cholesterol levels during statin therapy showed no correlation with the progression of aortic stenosis.17 Both Novaro and colleagues18 and Aronow and associates19 also reported similar results that statin-treated groups had a slower rate of aortic stenosis progression. Shavelle and coworkers9 have documented that patients receiving statins have a diminished rate of accumulation of aortic valve calcium, as assessed by means of electron beam computed tomography, although serum cholesterol levels were not available. It is an intriguing finding that the effects of the statins are not solely attributable to lipid-lowering effects. This is not without precedent because the statins have been shown to stabilize plaques as part of the beneficial effect in preventing cerebrovascular occlusion and in reducing inflammation. Currently, there is no information on the rate of progression of bioprosthetic valve sclerosis with statin therapy. In our case-control study patients whose valves were not explanted had a higher incidence of statin use than those whose valves were explanted (18% vs 6%, P = .018). Some patients might have been started on statins in the midst of their valve use. This prompted reanalysis with all statin pair groups removed. The mean serum cholesterol level was still higher (190.8 vs 158.5 mg/dL, P < .0001) when the confounding effects of statins were removed. We also matched patients so that operations occurred in the same year while keeping the age of the patient at implantation and position matched and found the explanted valve group to have a higher serum cholesterol level than the control subjects (187 vs 155 mg/dL, P = .047).
The correlation of increased cholesterol levels with bioprosthetic valvular calcification invites mechanistic speculation.5 Eliminating lipids from tissue valves by means of ethanol extraction prevents in vivo calcification.20 It is possible that glutaraldehyde-fixed valves have the appropriate physical characteristics to serve as a nidus for crystallization. Another potential mechanism is the migration of cells into the matrix of the valve. Blood-borne progenitor cells that are capable of differentiating into various lineages from the marrow might have the ability to deposit matrix, which can calcify. Additionally, oxidized LDL, which colocalizes with T lymphocytes and calcium, has been demonstrated in nonrheumatic stenotic aortic valves.21 Older literature had documented that silicone rubber prostheses in the Starr-Edwards valve (Edwards Lifesciences, Irvine, Calif) absorb lipids.22 Ferrans and associates23 had documented that cholesterol ester crystals were found in porcine aortic valves that had been implanted in the mitral position for 8 years.
This study has its limitations. It is retrospective, and the population for the cohort study is selected for patients who required reoperation. Patients who did not undergo reoperation were not included in the cohort study, although we attempt to address this in the case-control study. One could attempt to address this by using electron beam computed tomography to measure the progression of calcification over time in tissue valves. Serum cholesterol values were used rather than LDL levels. Smoking was self-reported.
In summary, we have shown in a retrospective cohort and case-control study that cholesterol might contribute to bioprosthetic valve calcification, which might lead to valve re-replacement. A randomized prospective clinical trial to determine whether statins prevent bioprosthetic valve calcium accumulation is warranted.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This paper was submitted early and actually has already been revised and accepted for publication in the Journal by Andrew Wechsler. Tirone David has written an editorial pertaining to this paper and a similar one from Germany by Nollert and associates concerning the effects of hyperlipidemia on the durability of the Hancock pericardial valve. Both papers will be fast-tracked because the Editor believes this question is important.
Tirone David looked at 635 old Hancock I valves from Stanford implanted in the 1970s and 1980s and observed for 15 years or more. He could show no adverse impact of hyperlipidemia on structural valve deterioration (SVD). He then looked at 913 patients undergoing aortic valve replacement at the Toronto General Hospital in the 1990s (Hancock II, Carpentier-Edwards pericardial, and Toronto SPV) and again could find no apparent adverse influence of hyperlipidemia on SVD.
Some of you may have been wondering why the freedom from SVD curves in Toronto with any kind of tissue valve are so good. Tirone himself has admitted that maybe it is because in Ontario almost everybody is already on a statin; maybe they never will see an effect there in their more recent patients. Another factor pertaining to the mitral position is that only the very sickest and very elderly patients undergo mitral valve replacement with a tissue bioprosthesis; the majority of valves are repaired.
In any event, this is promising work, even though it is just the first step. I do have one statistical question: When you get down to the case match controls, exclude the statin matches, and then you rematch for age, you get to some very small Ns. How does this affect the reliability of your statistics?
My real question is this: What is the next step? How are we going to confirm that this is true or not in larger, multicenter analyses? Tell us where you are headed. In today's world, at least in California, it appears that almost everybody is already on atorvastatin calcium (Lipitor) or some statin, even if they do not have coronary disease; therefore, how can you ever hope to test your hypothesis? What is the next step to obtain conclusive proof that treatment of hypercholesterolemia into the low normal range—and I should emphasize that the values shown here used to be accepted as normal in the North American population B—delays bioprosthetic SVD in large patient cohorts?
Dr Farivar. Dr Miller, thank you. This is the first report in bioprosthetic valves, and to get to the truth, I think we need a prospective randomized controlled trial. I would envision a multicenter trial enrolling subjects who have implanted tissue valves and subcritical coronary stenosis and are not receiving statins. We should randomize them into two groups—one to receive HMG-CoA reductase inhibitors and the other to receive placebo—and to follow the progression of bioprosthetic tissue valve calcium accumulation over 1 to 2 years, using methods that are similar to a study recently published concerning the native aortic valve. Shavelle and associates have recently shown in the Lancet that statins reduce the accumulation of calcium in the native aortic valve, as assessed by electron beam computed tomographic scanning, for calcium in the valve. These studies could be performed concurrently with surveillance scans of the coronary arteries to determine whether statins also prevent the progression of coronary calcification.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. Pibarot and J. G. Dumesnil Prosthetic Heart Valves: Selection of the Optimal Prosthesis and Long-Term Management Circulation, February 24, 2009; 119(7): 1034 - 1048. [Full Text] [PDF]
|
|
|
|
|
|
M. Briand, P. Pibarot, J.-P. Despres, P. Voisine, J. G. Dumesnil, F. Dagenais, and P. Mathieu Metabolic Syndrome Is Associated With Faster Degeneration of Bioprosthetic Valves Circulation, July 4, 2006; 114(1_suppl): I-512 - I-517. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Should all patients receive statins before cardiac surgery: are more data necessary? J. Thorac. Cardiovasc. Surg., March 1, 2006; 131(3): 520 - 522.
|
|
|
|
 |
|
 F. J. Schoen and R. J. Levy Calcification of Tissue Heart Valve Substitutes: Progress Toward Understanding and Prevention Ann. Thorac. Surg., March 1, 2005; 79(3): 1072 - 1080. [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 |
