Ann Thorac Surg 2004;77:1593-1597
© 2004 The Society of Thoracic Surgeons
Original article: cardiovascular
Effect of prior malignancy on survival after cardiac surgery
Wilhelm P. Mistiaen, MD, PhDa*,
Philip Van Cauwelaert, MDb,
Philip Muylaert, MDb,
Floris Wuyts, PhDc,
Fernand Harrisson, PhDa,
Hilde Bortier, MD, PhDa
a Laboratory for Anatomy, University of Antwerp, Antwerp, Belgium
b Department of Cardiovascular Surgery, General Hospital Middelheim, Antwerp, Belgium
c Medical Physics and Biometrics, University Hospital Antwerp, Edegem, Belgium
Accepted for publication August 28, 2003.
* Address reprint requests to Dr Mistiaen, Laboratory for Human Anatomy, University of Antwerp, Groenenborgerlaan 171, 2020 Antwerp, Belgium
e-mail: mistiaen{at}ruca.ua.ac.be
 |
Abstract
|
|---|
BACKGROUND: The number of patients with a previously treated tumor, needing cardiac surgery is increasing. Whether this operation in these patients is justified is determined by the long-term outcome.
METHODS: Of 8620 patients referred for cardiac surgery, 205 had a documented malignant tumor. The time interval between the occurrence of the tumor and the cardiac surgery was recorded. These patients were matched with 205 patients without a tumor according to age, gender, comorbidity and type of cardiac surgery. The patients were followed retrospectively. A 
2 Kaplan Meier and Cox' regression analysis were performed.
RESULTS: During follow-up, 95.8% of the patients were traced (2794 patient years). Univariate analysis showed that 5- and 10-year survival was better in patients without a malignant tumor in the history (0.91 ± 0.02 versus 0.72 ± 0.03 and 0.73 ± 0.04 versus 0.40 ± 0.05; p < 0.0001). For shorter time intervals, mortality for all causes and mortality due to the tumor increase significantly (p < 0.0001). Multivariate analysis identified 4 independent variables: a malignant tumor in the history (p < 0.001), chronic obstructive pulmonary disease (p = 0.003), age (p = 0.001), and impaired left ventricular function (p = 0.035)
CONCLUSIONS: A malignant tumor in the history is the most prognostic factor after cardiac surgery, but the operation is still rewarding. Fatal progression of the tumor is seen if the time interval between the occurrence of the malignant tumor and cardiac surgery is short. Other unfavorable factors are decreased left ventricular function, chronic obstructive pulmonary disease and high age.
|
Introduction
|
|---|
In an ageing population, the incidence of cardiovascular disease and of malignant tumors is increasing. With use of tobacco as a common risk factor for coronary heart disease and for several types of cancer [1–3], one could expect an increasing number of patients with a malignant tumor in the history also needing cardiac surgery. This is especially true because patients with increasing age and complex comorbidity are referred for cardiac surgery. Previous reports of the effect of such malignant tumor on survival after cardiac surgery were restricted mostly to patients with pulmonary tumors, small series, and short follow-up [4–9]. Several questions need to be addressed. How often do cardiologists and cardiac surgeons encounter these patients? Is a cardiac operation justified? What are the risk factors and what survival can be offered to these patients?
|
Patients and methods
|
|---|
During the period from January 1, 1989 to December 31, 1997, cardiac surgery with an extracorporeal circulation was performed upon 8620 patients. In a first step, the files of the patients who had left the hospital alive were studied. A malignant tumor in the history was documented in 205 patients. In a second step, these patients were matched with 205 patients without a tumor, according to age, gender, and type of cardiac surgery (number of coronary bypasses and valve surgery). In a third step, the degree of matching for comorbidity (diabetes, renal failure, chronic obstructive pulmonary disease) and left ventricular function was investigated. Almost all patients had surgery for severe coronary artery or valve disease with symptoms of New York Heart Association (NYHA) class III or IV.
The patients who had a malignant tumor in the history were treated for this disease in different centers, with different modes of surgery, chemotherapy, and radiotherapy. The intent of treatment was curative for all patients. Nevertheless, grading of the tumor could not always be obtained. Therefore, the time interval, expressed in years, between the occurrence of malignant tumor and cardiac surgery was included as a potential predictive factor for survival. If it was less than 1 year, or if a curable tumor was not yet treated at the time of cardiac surgery, it was assigned a value of 0. Patients that had an incurable tumor were excluded for cardiac surgery. For the most common organ sites of these malignant tumors, mortality was also determined.
The patients were followed retrospectively, using a questionnaire sent to cardiologists and general practitioners. This included symptoms, appearance of a metastasis or a new tumor, mortality (global, due to the malignant tumor and cardiac disease).
A Mann-Whitney U test and a 2 test were used to compare the preoperative profile of both patient groups. A Kaplan-Meier and Cox' proportional hazard analysis were used for the postoperative results and to identify the independent predictors for survival. A 2 analysis with linear to linear association was used to assess the effect of the time interval between the tumor and cardiac surgery on survival.
|
Results
|
|---|
Of the 205 patients with a malignant tumor in the history, 197 or 96.1% could be traced, with a total follow-up of 1243 patient years or 6.4 years per patient. Of the 205 patients without a malignant tumor, 196 or 95.6% could be traced, with a follow-up of 1551 patient years or 8.0 years per patient. After cardiac surgery, more than 90% of the patients were in NYHA class I or II. In 16 patients with a prior tumor, no time interval could be documented with certainty.
Preoperative profile
Table 1
illustrates the comparison of the preoperative data between patients with and without a documented malignant tumor. No significant differences could be observed between patients with and without a malignant tumor for any of the cofactors, nor for the type of surgery or the number of bypasses. We considered this as a reasonable match.
Postoperative mortality
Survival was significantly lower in patients with a malignant tumor in the history (p < 0.0001). After 5 years this is 0.91 ± 0.02 versus 0.72 ± 0.03, after 10 years, this was 0.73 ± 0.04 versus 0.40 ± 0.05.
Figure 1
demonstrates the effect of the time interval between the occurrence of the tumor and the cardiac surgery on the mortality for all causes. Figure 2
reveals the effect of this time interval on mortality due to the tumor. Patients without a malignant tumor in the history had the highest survival. Mortality increased with shorter time intervals, especially if this was less than 2 years. This effect is highly significant (p < 0.0001) on mortality for all causes as well as on mortality due to the tumor. This was especially evident in patients with simultaneous cardiac bypass operation and pulmonary surgery for a tumor (n = 8) or in whom surgery for a curable tumor had to be postponed until after cardiac surgery (n = 3). Nine of these 11 patients died during follow-up, 8 of them due to a tumor progression. Their median survival time was 2.0 years (1.3 to 5.4 years).
View larger version (20K):
[in this window]
[in a new window]
|
Fig 1. Illustration of mortality for all causes: patients without a tumor in the history (——), for patients with an interval between the occurrence of the tumor and cardiac surgery of more than 10 years (· · ·), between 6 and 10 years ( – – – ), between 2 and 5 years (– · · –), and less than 2 years (— — —). Patients at risk at 0 (at discharge), at 5 years and at 10 years: no tumor: 195, 182 and 63; interval more than 10 years: 47, 38 and 11; interval 5 to 10 years: 37, 29 and 5; interval 2 to 5 years: 53, 40 and 9; interval less than 2 years: 44, 24 and 3.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig 2. Mortality due to tumor: patients without a tumor in the history (——), for patients with an interval between the occurrence of the tumor and cardiac surgery of more than 10 years ( · · · ), between 6 and 10 years (– – –), between 2 and 5 years (– · · –), and less than 2 years (— — —). Patients at risk at 0 (at discharge), at 5 years and at 10 years: no tumor: 195, 182 and 63; interval more than 10 years: 47, 38 and 11; interval 5 to 10 years: 37, 29 and 5; interval 2 to 5 years: 53, 40 and 9; interval less than 2 years: 44, 24 and 3.
|
|
Mortality depended also on the origin of the tumor as can be seen in Table 2,
which illustrates the distribution and the mortality of the most common tumors in the cohort. Other tumors were less common. High mortality in pulmonary carcinoma was related to the short time interval between the occurrence of the tumor and the cardiac surgery. Lethality in prostate cancer was also related to the time interval: only 4 of 28 patients had an interval between the tumor and the cardiac surgery of more than 5 years. Most other common tumors had a mortality of less than 50% and 53.4% of these patients had a time interval of more than 5 years.
Postoperative mortality according to the cause
Table 3
indicates the cause of death in relation to the time interval: in patients with a prior tumor, the postoperative death rate due to the tumor increased with decreasing time interval. A postoperative recurrence of the preoperative tumor was observed in 44 patients. Despite chemotherapy and radiotherapy, 36 patients died. In 34, this was due to widespread metastasis (n = 31) or local recurrence (n = 3). The remaining 2 patients with recurrence died due to pneumonia or with dementia.
In patients with a prior malignant tumor, a second and histologic different primary tumor occurred in 25 patients (12.8%). Twelve patients died, of which 10 from this new tumor, at 60 months (range 22–84 months) after operation. One of these patients had also a metastasis from the preoperative tumor. In 20 patients (10.2%) from the group without prior malignant tumor, a tumor appeared during the follow-up. Ten patients died from this disease, at 69 months (range 38–93 months) after cardiac surgery. The difference between both groups is not significant (p = 0.45).
Identification of independent predictors by multivariate analysis
Table 4
identifies four independent factors by a Cox proportional hazard analysis. The presence of a malignant tumor in the history was the most important factor. Diabetes, renal failure, arterial hypertension, type of cardiac surgery, and the number of bypasses had no significant effect.
|
Comment
|
|---|
About every 2 weeks a patient was referred for cardiac surgery who had a history of a previously treated or a currently curable malignant tumor. This could largely be attributed to a use of tobacco, which is a common risk factor for coronary artery disease as well as for several forms of malignant tumor [1–3].
In a previous report, we found in 400 patients undergoing aortic valve replacement with a pericardial valve, a hospital mortality of 3.8%. In 38 patients with a previous malignant tumor, this was 2.6%, which was not significantly different [10]. Because there was no obvious reason for increased hospital mortality after cardiac surgery in patients with a previous or current malignant tumor, we focused the investigation on long-term survival.
Our analysis revealed that the presence of a malignant tumor was the dominant independent predictor for survival after cardiac surgery. This was followed by a chronic obstructive pulmonary disease, impaired left ventricular function, and age. The other investigated factors had no significant effect on the survival after cardiac surgery.
The mortality of patients with a malignant tumor was mostly due to the progression of this disease: the tumor was responsible for almost 50% of the fatalities and the survival time in these patients was much shorter.
Use of grading of the tumor as data to study the effect on survival had disadvantages. The information was only available in a minority of patients, mostly in those with a small time interval between the occurrence of the tumor and the cardiac surgery. This could induce a bias. Moreover the mode of treatment differed between the referring hospitals, making comparison between patients difficult. Therefore the time interval between the previous malignant tumor and the cardiac surgery was preferred as a variable. This interval could be seen as a rough measure for the probability for cure from the malignant tumor after treatment with curative intent. This factor had a major effect on the survival after cardiac surgery. In short time intervals, mortality was high and mostly due to the progression of the malignant tumor. For longer intervals, mortality was lower and more due to other causes. This effect was striking in 11 patients with no time interval.
Previous data from several investigators [11–14] suggests that the use of a cardiopulmonary bypass induced an inflammatory reaction, including an increased release of cytokines (interleukin-1ß, interleukin-6, interleukin-8, and tumor necrosis factor-
). An antiinflammatory response has also been recorded, with release of interleukin-10 and a decrease in activity of interleukin-2 and interleukin-12, which are of importance in the regulation of cell-mediated immunity. After cardiac surgery, a shift of T-cell phenotype toward humoral immunity has been observed, with a concomitant suppression of cell mediated immunity. T cells with a CD4+ expression were also decreased, up to 7 days after the operation.
Some of these cytokines had also been involved in the immunologic defense against tumor cell formation, invasion, and metastasis. In general, a central role of suppressed cell mediated immunity in the pathogenesis of malignant diseases has been observed [15]. Interleukin-2 and interleukin-12, as well as interferon- and interferon-
stimulated antitumor immunity or prevented the onset of genetically determined tumors, whereas interleukin-6 had immunosuppressive activity by counteracting the effect of interleukin-2. Suppression of CD4+ activity on T cells has also been observed in immunosuppression regimen and carcinogenesis, in some instances by a decrease of tumor surveillance activity [16–19].
A common role of interleukins could be assumed in suppression of immunity after use of a cardiopulmonary bypass as well as in carcinogenesis. Nevertheless, there were three reasons to exclude the use of a cardiopulmonary bypass as a causative mechanism.
First, changes in the immune system after cardiac surgery with a cardiopulmonary bypass were only investigated during the hospital stay and were probably no longer than 7 days. In contrast, the effect of factors such as pollutants and use of tobacco on the immune system was chronic. Therefore, a short alteration of the immune system was not likely to induce carcinogenesis, unless it could trigger a progression of a dormant malignancy.
Second, after cardiac surgery, we did not find an increase of a new primary tumor in patients with a previous carcinoma, compared to patients without a previous tumor.
Third, we found a profound effect of the length of the time interval between the occurrence of a malignant tumor in the history and cardiac surgery on postoperative mortality. Since this interval could be regarded as a measure for the probability of cure, this observation indicates that mortality after cardiac surgery in patients with a previously treated malignant tumor was mostly due to the tumor progression. This interval can therefore be used in the therapeutic decision process of cardiac surgeons, because, in contrast to the staging and the precise nature of treatment of the prior tumor, it is readily available and has a clear impact.
Although the current results are predictable, they can affect therapeutic decisions in these patients. On the one hand, several symptom-free years could be obtained after cardiac surgery, which is rewarding. On the other hand, severe cardiac symptoms can persist without cardiac surgery. However, patients with a malignant tumor treated less than 2 years earlier, high age, pulmonary disease, or impaired left ventricular function should be made aware of the limited survival. A decision to perform cardiac surgery in patients with a coincidentally found malignant tumor, which is potentially curable, must be made with care.
|
References
|
|---|
- Kuper H., Bofetta P., Adami H.O. Tobacco use and cancer causation: association by tumor type. J Intern Med 2002;252:204-206.
- Fagerstrom K. The epidemiology of smoking: health consequences and benefits of cessation. Drugs 2002;62(Suppl 2):1-19.
- Thomas L.A. Sending unhealthy environments up in smoke. Mich Health Hosp 2002;38:20-21.
- Danton M.H.D., Anikin V.A., McManus K.G., McGuigan J.A., Campalani G. Simultaneous cardiac surgery with pulmonary resection: presentation of series and review of literature. Eur J Cardiothorac Surg 1998;13:667-672.
- Al-Attar N., Salvi S., Sebbag U., Nataf P. Combined left pneumonectomy and off-pump coronary artery bypass through left thoracotomy. Eur J Cardiothorac Surg 2001;19:226-228.[Abstract/Free Full Text]
- Hensens A.G., Zeebregts C.J., Liem T.H., Gehlmann H., Lacquet L.H. Concomitant coronary artery revascularisation and right pneumonectomy without cardiopulmonary bypass. J Cardiovasc Surg 1999;40:161-163.[Medline]
- Rubens F.D., O'Brien E., Shamji F., O'Neil J., Odell P.F. Sternothoractomy for combined coronary artery bypass grafting and left upper lobectomy in a patient with low-lying tracheostoma. Can J Surg 1999;42:143-144.[Medline]
- Jones J.M., Melua A.A., Anikin V., Campalini G. Simultaneous transhiatal esophagectomy and coronary artery bypass grafting without cardiopulmonary bypass. Dis Esophagus 1999;12:312-313.[Medline]
- Yokoyama T., Derrick M.J., Lee A.W. Cardiac operation with associated pulmonary resection. J Thorac Cardiovasc Surg 1993;105:912-917.[Abstract]
- Mistiaen W.P., Van Cauwelaert P., Muylaert P., et al. Determinants of survival after aortic valve replacement as treatment for symptomatic aortic valve disease in the elderly. J Heart Valve Dis 2001;10:354-360.[Medline]
- Markewitz A., Lante W., Franke A., Marohl K., Kuhlmann W.D., Weinhold C. Alterations of cell-mediated immunity following cardiac operations: clinical implications and open questions. Shock 2001;16(Suppl 1):10-15.
- Sablotski A., Welters I., Lehmann N., et al. Plasma levels of immunoinhibitory cytokines interleukin-10 and transforming growth factor-beta in patients undergoing coronary artery bypass grafing. Eur J Cardiothorac Surg 1997;11:763-768.[Abstract]
- Markewitz A., Faist E., Lang S., Hultner L., Weinhold C., Reichart B. An imbalance in T-helper cells subsets alters immune response after cardiac surgery. Eur J Cardiothorac Surg 1996;10:61-67.[Abstract]
- Payen D., Faivre V., Lukaszewicz A.C., Losser M.R. Assessment of immunological status in the critically ill. Minerva Anestesiol 2000;66:757-763.[Medline]
- O'Byrne K.J., Dalgleish A.G., Browning M.J., Steward W.P., Harris A.L. The relationship between angiogenesis and the immune response in carcinogenesis and the progression of malignant disease. Eur J Cancer 2000;36:151-169.
- Nanni P., Nicoletti G., De Giovanni C., et al. Combined allogeneic tumour cell vaccination and systemic interleukin 12 prevents mammary carcinogenesis in HER-2/leu trangene mice. J Exp Med 2001;194:1195-1205.[Abstract/Free Full Text]
- Lamm D.L., Riggs D.R. The potential application of Allium Sativum for the treatment of bladder cancer. Urol Clin North Am 2000;27:157-162.[Medline]
- Mise-Omata S., Sugiura T., Higashi K., Yamashita U. Reduces T cell response in carcinogen-sensitive Donryu rats compared with carcinogen-resistant DHR rats. Jpn J Cancer Res 1999;90:1277-1284.[Medline]
- Enwonwu C.O., Meeks V.I. Bionutrition and oral cancer in humans. Crit Rev Oral Biol Med 1995;6:5-17.[Abstract/Free Full Text]
Top
Abstract
Introduction
