| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Division of Cardiac Surgery, Ospedali Riuniti di Trieste, Trieste, Italy
Accepted for publication October 8, 2008.
* Address correspondence to Dr Zingone, 22, Vicolo Scaglioni, Trieste, 34141, Italy (Email: bartolo.zingone{at}aots.sanita.fvg.it).
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentReferences |
|---|
Methods: Records of consecutive patients 80 years old or greater (n = 355) having cardiac operations from September 1998 through May 2007 were reviewed. There were 172 isolated coronary bypass grafting (CABG), 73 isolated valve, 79 valve and CABG combined, and 31 other procedures.
Results: Thirty-three (9.3%) deaths and 13 (3.7%) strokes occurred during the index hospital stay. Intensive care unit and hospital length of stay lasted 6.3 ± 14.3 and 15.5 ± 20.8 days, respectively. Overall cumulative 5-year survival was 65.5 ± 3.3%, varying among procedures as follows: 67.9 ± 4.4% for isolated CABG, 64.6 ± 8.9% for valve surgery, 60.3 ± 7.3% for combined coronary and valve surgery, and 63 ± 10.7% for other procedures (p = 0.23). Ninety-seven percent of survivors lived at home. Risk factors for hospital death were emergency status, preoperative renal dysfunction, and postoperative complications such as myocardial infarction, cardiac failure requiring intraaortic balloon pumping, acute renal failure requiring replacement therapy, stroke, and ventilator dependency exceeding 48 hours. Among hospital survivors, risk factors for late death were carotid artery disease, chronic lung disease, renal dysfunction, and the occurrence of postoperative complications.
Conclusions: Long-term survival of octogenarians submitted to a wide variety of cardiac operations was satisfactory despite substantial rates of early complications and deaths. Most survivors were free from cardiac symptoms. Postoperative complications were stronger risk factors for hospital deaths than preoperative comorbidities and procedural variables. Their impact on long-term survival was also significant.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
On the other hand, a number of follow-up studies on operated patients show encouraging midterm survival and quality of life [5–9]. Admittedly, these studies are heterogeneous in terms of subtended eras and selection processes, often focusing on specific procedures and techniques, and might be misrepresentative due to publication bias. In addition, their reproducibility remains open to question as long as good results are mainly reported from leading centers. We therefore reviewed our own series of consecutive octogenarians and the full spectrum of surgical procedures in order to assess the value of the treatment and to identify areas for improvement.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Surgery was carried out for the majority of patients on cardiopulmonary bypass with hollow-fiber oxygenators and cold blood cardioplegia. Core temperature was allowed to drift to 32°C to 34°C except for cases requiring deep hypothermic circulatory arrest or off-pump coronary artery bypass grafting (OPCABG). Intraoperative management included ultrasonographic scanning of the ascending aorta in most patients and transesophageal echocardiography for valve repairs.
Definitions of preoperative comorbidities were those employed for the European System for Cardiac Operative Risk Evaluation (EuroSCORE) [10]. Creatinine clearance was estimated by a simplified Cockcroft-Gault formula [11] in which measured body weight substituted lean body weight. Calcified or atheromatous aortas were assessed preoperatively by plain chest X-ray or computed tomographic scanning, and intraoperatively by ultrasonographic techniques.
The main determinant of intensive care unit stay longer than two days was prospectively singled out of a list of 11 and recorded. For the purpose of this study, it was also taken as the cause of hospital death at time of data retrieval, except for patients dying within two days whose deaths were adjudicated retrospectively. Hospital deaths were those occurring any time before hospital discharge.
Complications of interest, independent from the list mentioned above, were declared before data retrieval. They were selected for their potential for early occurrence and their expected impact on further hospital course. Postoperative renal failure was diagnosed if peak postoperative creatinine was 1.5 or greater times the preoperative value. Stroke was defined as a new focal neurologic deficit or coma appearing during the postoperative stay, and at least partially evident more than 24 hours after onset.
Follow-up was obtained during a two-month closing interval ending November 2007 by telephone interview with the patients or their relatives. The vital status of untraced patients was ascertained through the regional health statistical office ("Servizio Informatico Applicazioni Sanitarie Integrate"). No patient was lost to follow-up.
Categoric data are expressed as counts and proportions, and differences between groups assessed by the Pearson 
2 or Fisher exact test as appropriate. Continuous variables are expressed as mean ± standard deviation and compared for differences between groups with the Mann-Whitney U test. Expected in-hospital death rates were estimated by the logistic EuroSCORE [10]. Variables potentially associated with in-hospital death were assessed by multiple logistic regression analysis and their odds ratios (OR) reported with 95% confidence intervals (CI). Models discrimination was assessed by receiver operating characteristic (ROC) analysis. Kaplan-Meier survival curves included hospital deaths and were compared by the log-rank test. The age- and sex-matched survival curve of the Italian population was plotted for reference [12]. Factors univariately associated with events of interest (p 
0.05) were entered by a stepwise forward inclusion method into a number of Cox proportional hazard models. Variables of special interest were forced into the models irrespective of their univariate significance. Multiple sets of variables were used to develop different models in the entire cohort and, separately, in the hospital survivors cohort. Propensity scores for contrast variables were developed by nonparsimonious modeling with preoperative data and adjusted for in the appropriate models. Hazard ratios are reported together with their 95% CI. All tests were 2-tailed. Differences were considered significant at a p value 0.05 or less; SPSS 14.0 (SPSS Inc, Chicago, IL) was used throughout.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
|
Early Outcomes
Reoperations were required for bleeding or tamponade (n = 18), sternal rewiring (n = 10) valve repair failure (n = 1), myocardial infarction (n = 1), and mediastinitis (n = 1). Five patients died after reoperation for bleeding (n = 4) or mediastinitis (n = 1).
Death and complication rates are shown in Table 2. Compared with EuroSCORE based mortality estimates, observed death rates were significantly lower for the whole cohort and for some of the main procedural subgroups. Causes of death were multiple organ failure (n = 14), cardiac failure (n = 9), respiratory failure (n = 4), bleeding (n = 2), neurologic damage (n = 2), bowel ischemia (n = 1), and renal failure (n = 1).
|
|
|
A further regression was then run with preoperative variables, a propensity score for postoperative complications, and with the individual complications themselves. While emergency status (OR, 6.45; CI, 1.61 to 25.90) and preoperative serum creatinine (OR, 2.48; CI, 1.45 to 4.23) retained their significance, age (OR, 1.29; CI, 0.99 to 1.69) did not. In addition, postoperative myocardial infarction (OR, 18.32; CI, 3.40 to 98.61), acute renal failure requiring replacement therapy (OR, 15.89; CI, 2.88 to 87.67), cardiac failure requiring intraaortic balloon pumping (IABP) (OR, 11.81; CI, 1.52 to 91.63), prolonged intubation (OR, 11.42; CI, 3.53 to 37.02), and stroke (OR, 5.38; CI, 1.11 to 26.10) were independently associated with the risk of hospital death. Areas under the ROC curve were 0.65 for the logistic EuroSCORE alone, 0.67 for the model based on preoperative variables only, 0.69 for the preoperative set adjusted for EuroSCORE, and 0.94 after the addition of postoperative complications to the preoperative set of variables.
The predictability of complications independently associated with death was also explored. Despite the emergence of peripheral artery disease (OR, 2.15; CI, 1.23 to 3.74) and urgent status (OR, 1.84; CI, 1.07 to 3.17) as risk factors in addition to EuroSCORE (OR, 1.03; CI 1.01 to 1.05) and creatinine clearance (OR, 0.97; CI, 0.95 to 0.99), the ROC area for this model remained at 0.67.
Late Outcomes
The cause of death could be adjudicated in 48 of 67 late events as cardiac related (n = 14, including 5 sudden unexplained events), neurologic events (n = 13), cancer (n = 11), respiratory failure (n = 4), bowel infarct (n = 2), and diabetes, road accident, renal failure, and suicide (n = 1 each). Overall crude survival (± standard error) at 1, 3, and 5 years was 83.5 ± 2.0%, 73.8 ± 2.6%, and 65.5 ± 3.3%, respectively, inclusive of hospital events. Five-year survival (Fig 1A) was 67.9 ± 4.4% for isolated CABG, 64.6 ± 8.9% for isolated valve surgery, 60.3 ± 7.3% for combined CABG and valve surgery, and 63.0 ± 10.7% for the remaining cases (p = 0.23).
|
Table 5 enlists factors independently associated with death for models 1 and 2 and shows the unfavorable effect of complications on overall survival. The analysis was then repeated for patients discharged alive only, but the assumption that the effect of complications may extinguish after the acute perioperative phase proved to be wrong. A continuing attrition over time was in fact demonstrated (Table 5, Fig 2).
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Hospital mortality rate in our series was well within the narrow range of 7.5 to 9.8% of similar all-inclusive studies [5, 6, 13]. Greater variability can be found among studies with different exclusion criteria [7–9, 14–19], with rates as low as 1.7% after isolated mitral valve surgery for degenerative disease [20] and as high as 19.6% for combined mitral and CABG surgery [15]. Although these figures make preoperative risk stratification desirable, this has been reported only occasionally in this age subset. Stoica and colleagues [6] found that EuroSCORE may significantly overestimate the operative risk in octogenarians. Although previously validated in our general population [21], and in keeping with the findings of Stoica and colleagues, EuroSCORE would have predicted a 76% higher risk of mortality than observed in our octogenarians. Relying on an inadequately calibrated model, therefore, may be misleading and further encourage a conservative attitude when selecting treatment for aged individuals. On occasion, it may let patients be inappropriately shifted across competing alternatives.
Preoperative comorbidities identified as risk factors in our own series were not helpful in developing a better explanatory model. Inadequate power may have contributed, but the issue has also been raised in a larger series before [13]. That no risk factor constituted a criterion to possibly contraindicate surgery is best exemplified by surgical priority, a long-held major predictor for operative death. While emergency status turned out to be significantly associated with death by multivariable analysis, 74% of emergency cases were discharged alive, a likely better outlook than the expected alternative. Furthermore, the discriminating ability of the models including age, renal dysfunction, and emergency status together was associated with c-statistics disappointingly lower than 0.7. Finally, when adjusted for comorbidities and surgical factors, age greater than 80 maintains a residual explanatory power presumably due to other unmeasured characteristics that have not yet been defined [19, 22].
It seems therefore clear that improving the discrimination of patients less likely to benefit from surgery will require expanding the array of candidate predictors to include, among others, specific measures of physical and mental functioning in the elderly [23]. For instance, Oxman and colleagues [24] found that, in addition to age and history of previous surgery, impairment of basic daily life activities, lack of social participation, and religious comfort were independently associated with 6-month mortality. Rumsfeld and colleagues [25] demonstrated that health-related quality of life measures, and more specifically of physical functioning, were significant predictors of 6-month mortality even after adjusting for traditional variables. Frailty, a prominent criterion when selecting elderly patients for surgery, is usually assessed in a subjective manner, although in broader settings [26] it could be quantified by a number of indicators that should probably be incorporated into both clinical practice and future studies. Still, while the search for an objective way to identify those patients who are poor surgical candidates is strongly needed, it should be noted that a subjective approach raising the octogenarians' fraction of our yearly caseload from 4.2% to 14.7% during the study period has caused no increase in the operative mortality rates.
Postoperative complications have been recorded in approximately half the cases in previous series focusing on elderly patients [13, 27]. We obtained a slightly lower rate of complications by limiting the analysis to those occurring at a very early stage after surgery. These more likely dictate the further course of events, and in fact effectively discriminated patients bound to have a longer hospital stay, also capturing almost all fatal hospital events. When included in the search for risk factors for hospital death they significantly increased the explanatory power of the regression model. Finally, their unfavorable influence on survival persisted over the years and was confirmed by the multivariable analysis on hospital survivors. These observations, if combined with the limited accuracy of preoperative stratification discussed before, point to the importance of operative events and surgical factors in determining postoperative morbidity and fatalities not only in the hospital [13, 16] but also in the long term [16]. Flawless surgery, meticulous hemostasis, perfect myocardial protection, and adequate organ perfusion, however, are basic requirements for everyday practice. The degree they are implemented, on the other hand, may make the difference as the threshold for entering a complicated course may be lower and that for getting out of it higher for the elderly.
In regard to technical aspects and their relation to outcome, our series was too small and heterogeneous to support any robust inference. It is notable, however, that times on cardiopulmonary bypass and duration of cardioplegia were univariately associated with increasing rates of complications but did not affect survival. Procedures requiring hypothermic circulatory arrest were also reasonably well tolerated. These were unexpected findings, at variance with previous reports [6, 13] and conventional wisdom, for which we have no better explanation than chance alone. However, although we would definitely not endorse disregarding the time allocated to the procedures, policies such as extensive arterial grafting, selective use of off-pump revascularization, and radically fixing all that is needed may convey more benefits than hitherto thought of.
The attention that early outcomes deserve should not distract from appreciating the value of long-term survival and quality of life. Two-thirds of patients at risk survived through their fifth postoperative year, resembling the outlook of the "normal" matched population and likely outliving, in aggregate, their expectancy without the operation. Most surviving patients also enjoyed full freedom from cardiac limitations and were able to live alone or with their relatives at home. Our concerns for occasional difficulties in getting hold of them often dissipated at realizing they were so busy with outdoor activities. Similar findings have been consistently reported before [5–9, 16–18, 20, 28] and suggest improved survivorship over the natural history of multiple diseases. In addition, evidence of a survival advantage for operated patients with aortic stenosis [29, 30] and coronary disease [31] has recently become available.
Our study suffers all the limitations of retrospective observations in highly selected patients. Furthermore, quality of life was not formally assessed in survivors, and escaped any evaluation in patients found to be dead at follow-up. Although it might have been even worse without operation that remains partly speculative and can only be clarified by a prospective study. The apparently limited impact of complex surgery on survival might be due to the small number of patients. Finally, the attrition of postoperative complications on late survival is disturbing but cannot be explained by our data. These limitations should be carefully considered while interpreting our results.
Survival of octogenarians after cardiac procedures of variable complexity, despite a substantial hospital mortality rate, resembles that of the normal aged individuals and is likely better than permitted by their disease. Survivors enjoy persisting freedom from cardiac symptoms and lead an independent, active life. Cautiously expanding surgical indications is largely based on empirical assessment and can be successful, although the limit it can be pursued and the criteria for patient selection remain to be defined.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
80 years: results from the National Cardiovascular Network J Am Coll Cardiol 2000;35:731-738.[Medline]80 years Eur J Cardiothorac Surg 2006;30:722-727.Related Article
This article has been cited by other articles:
|
|
 |
|
  L. Henry, L. Halpin, S. Hunt, S. D. Holmes, and N. Ad Patient Disposition and Long-Term Outcomes After Valve Surgery in Octogenarians Ann. Thorac. Surg., September 1, 2012; 94(3): 744 - 750. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. AlWaqfi, Y. Khader, and K. Ibrahim Coronary artery bypass: predictors of 30-day operative mortality in Jordanians Asian Cardiovascular and Thoracic Annals, June 1, 2012; 20(3): 245 - 251. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Mohammadi, D. Kalavrouziotis, F. Dagenais, P. Voisine, and E. Charbonneau Completeness of Revascularization and Survival Among Octogenarians With Triple-Vessel Disease Ann. Thorac. Surg., May 1, 2012; 93(5): 1432 - 1437. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Li, E. A. Amsterdam, K. K. Yeo, H. Hoegh, G. Mahendra, W. D. Boyd, and J. N. Young Coronary Artery Bypass Operations for Elderly Patients in California, 2003 to 2008 Ann. Thorac. Surg., April 1, 2012; 93(4): 1167 - 1172. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Afilalo, S. Mottillo, M. J. Eisenberg, K. P. Alexander, N. Noiseux, L. P. Perrault, J.-F. Morin, Y. Langlois, S. M. Ohayon, J. Monette, et al. Addition of Frailty and Disability to Cardiac Surgery Risk Scores Identifies Elderly Patients at High Risk of Mortality or Major Morbidity Circ Cardiovasc Qual Outcomes, March 1, 2012; 5(2): 222 - 228. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. S. Cooper, J. R. Charpie, F. X. Flores, J. William Gaynor, J. W. Salvin, P. Devarajan, and C. D. Krawczeski Acute Kidney Injury and Critical Cardiac Disease World Journal for Pediatric and Congenital Heart Surgery, July 1, 2011; 2(3): 411 - 423. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Saito, N. Motomura, H. Miyata, S. Takamoto, S. Kyo, M. Ono, and Japan Cardiovascular Surgery Database Organization Age-specific risk stratification in 13488 isolated coronary artery bypass grafting procedures Interact CardioVasc Thorac Surg, April 1, 2011; 12(4): 575 - 581. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Giallauria, C. Vigorito, R. Tramarin, F. Fattirolli, M. Ambrosetti, S. De Feo, R. Griffo, C. Riccio, and on behalf of the ISYDE-2008 Investigators of the I Cardiac Rehabilitation in Very Old Patients: Data From the Italian Survey on Cardiac Rehabilitation-2008 (ISYDE-2008)--Official Report of the Italian Association for Cardiovascular Prevention, Rehabilitation, and Epidemiology J Gerontol A Biol Sci Med Sci, December 1, 2010; 65A(12): 1353 - 1361. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Maraschini, F. Seccareccia, P. D'Errigo, S. Rosato, G. Badoni, G. Casali, and F. Musumeci Role of gender and age on early mortality after coronary artery bypass graft in different hospitals: data from a national administrative database Interact CardioVasc Thorac Surg, November 1, 2010; 11(5): 537 - 542. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Katlic Consider surgery for elderly patients Can. Med. Assoc. J., September 21, 2010; 182(13): 1403 - 1404. [Full Text] [PDF]
|
|
|
|
|
|
F. G. Bakaeen, D. Chu, J. Huh, and B. A. Carabello Is an Age of 80 Years or Greater an Important Predictor of Short-Term Outcomes of Isolated Aortic Valve Replacement in Veterans? Ann. Thorac. Surg., September 1, 2010; 90(3): 769 - 774. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. K. Ghanta, P. S. Shekar, S. McGurk, D. M. Rosborough, and S. F. Aranki Nonelective cardiac surgery in the elderly: Is it justified? J. Thorac. Cardiovasc. Surg., July 1, 2010; 140(1): 103 - 109. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Aiyagari, S. Gelehrter, E. L. Bove, R. G. Ohye, E. J. Devaney, J. C. Hirsch, J. G. Gurney, and J. R. Charpie Effects of N-acetylcysteine on renal dysfunction in neonates undergoing the arterial switch operation J. Thorac. Cardiovasc. Surg., April 1, 2010; 139(4): 956 - 961. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Nissinen, J.-O. Wistbacka, P. Loponen, K. Korpilahti, K. Teittinen, M. Virkkila, M. Tarkka, and F. Biancari Coronary Artery Bypass Surgery in Octogenarians: Long-Term Outcome Can Be Better Than Expected Ann. Thorac. Surg., April 1, 2010; 89(4): 1119 - 1124. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Toole, J. S. Ikonomidis, W. Y. Szeto, J. L. Zellner, J. Mulcahy, R. L. Deardorff, and F. G. Spinale Selective endothelin-1 receptor type A inhibition in subjects undergoing cardiac surgery with preexisting left ventricular dysfunction: Influence on early postoperative hemodynamics J. Thorac. Cardiovasc. Surg., March 1, 2010; 139(3): 646 - 654. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. B. Ross Invited Commentary Ann. Thorac. Surg., January 1, 2009; 87(1): 78 - 78. [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 |
), isolated coronary artery bypass grafting (CABG; 
), combined valve + CABG procedures (
), and other procedures (•). Survival of age- and sex-matched Italian population plotted for reference (straight line).