|
|
||||||||
a Department of Cardiothoracic Surgery, Weill Cornell Medical College, New York, New York
b Center for Complementary and Integrative Medicine, Weill Cornell Medical College, New York, New York
c Department of Statistical Science, Cornell University, Ithaca, New York
Accepted for publication October 23, 2007.
* Address correspondence to Dr Peterson, Center for Complementary and Integrative Medicine, Weill Cornell Medical College, 1300 York Ave, Box 46, New York, NY 10021 (Email: jcpeters{at}med.cornell.edu).
| Abstract |
|---|
|
|
|---|
Methods: A consecutive series of nonagenarians who underwent cardiac operations between 1995 and 2004 were retrospectively reviewed. Data collection included baseline preoperative clinical status, intraoperative characteristics, and perioperative course. Area under the Kaplan-Meier survival estimate method was used to calculate mean survival.
Results: Cardiac surgical procedures were done in 49 patients (51% male); their mean age was 91.9 years (range, 90 to 97 years). Operative mortality was 8% (n = 4). Multivariate Cox proportional hazards models found preoperative chronic renal insufficiency (hazard ratio [HR], 4.88; 95% confidence interval [CI], 1.53 to 15.55; p = 0.007) and ejection fraction (HR, 0.96; 95% CI, 0.93 to 1.00; p = 0.033) were independently associated with death. Overall mean survival was 5.1 ± 0.5 years (median, 5.2 years). Quality of life outcomes were similar to that of two related norm-based populations based on age and disease process.
Conclusions: Cardiac surgical procedures can be performed safely and with therapeutic benefit in carefully selected nonagenarians. We consider physiologic indicators, social factors, and patient preferences to be the main determinants in the patient selection process. Our results support the need for more proactive intervention in symptomatic nonagenarian patients as it relates to earlier consideration of elective, rather than emergency cardiac operations.
| Introduction |
|---|
|
|
|---|
Reports targeting cardiac operations in nonagenarians are limited. The current data are insufficient to adequately assess early to mid-term outcomes of cardiac surgical intervention and guide clinical practice. In addition, a variety of statistical methods and survival definitions have been used in reporting survival outcomes in previous reports, leading to an inability to compare results across the few existing studies in this area. This study reports a large series of nonagenarians at a single institution and adds to the literature by reporting a long follow-up period. In addition, the current study reports quality of life outcomes on a subset of survivors. We aim to advance existing literature by offering a long-term, complete statistical portrait of survival outcomes in nonagenarians. Furthermore, we aim to expand on previous efforts related to quality of life outcomes in this population.
| Material and Methods |
|---|
|
|
|---|
A retrospective analysis was performed using chart extraction to assess preoperative clinical profile, intraoperative characteristics, and postoperative status. Presence of coronary risk factors and comorbidities were recorded, and the Charlson comorbidity index [17] was calculated. Operative data were collected, including type of surgical procedure, bypass and cross-clamp times, and mean arterial pressure (MAP) during bypass. The perioperative and postoperative course was followed up for the occurrence of complications or death, transfusion requirements, use of inotropes and vasopressors, length of stay (LOS), and discharge plans.
The current disposition of all patients was reassessed in 2005 through chart review, or communication with the patient or the patients primary care physician, cardiologist, or family members. The Social Security Death Index [18] was used to locate lost patients. For survivors, the Medical Outcomes Study Short Form 12 (SF-12) Health Survey, version 2 [19] was used to assess quality of life.
Statistical Analysis
All data were imported and analyzed using R [20], Stata 9 [21], and SAS 8e software [22]. For all analyses, statistical significance was set at a p < 0.05. Differences in proportions were analyzed using
2 and the Fisher exact test. Mean differences for continuous variables were compared using Student t test. A Bonferroni correction was used to adjust for multiple comparisons. A multivariate Cox proportional hazards regression analysis Akaike information criterion (AIC) model selection procedure [23] was undertaken to identify independent predictors for survival. Survival curves were calculated using the Kaplan-Meier method to determine survival outcomes for the cohort.
Hazard ratios with corresponding 95% confidence intervals (CI) and p values were calculated for univariate Cox regression models. Each of these analyses was stratified by gender and robust standard errors were used [24]. A forced AIC model selection procedure [25] was used to select the three best fitting covariates for the multivariate model. In addition, robustness of the inferential results for the three variable Cox model was assessed using a bootstrap methodology [26]. All of the Cox models were validated using Schoenfeld residuals [24]. To account for censoring, the area under the Kaplan-Meier survival estimate method [27] was used to calculate mean survival, and the median survival was calculated from the 50th percentile of the Kaplan-Meier survival estimate.
| Results |
|---|
|
|
|---|
|
Cardiopulmonary bypass (CPB) was used for 47 operations, and two CABG procedures were performed off-pump. One-third of cases were emergencies, consisting of nine CABG, one AVR, three CABG/AVR, two ascending aortic aneurysm repairs, and one CABG/ascending aortic aneurysm repair. The average bypass and cross-clamp times were 94.4 ± 36 minutes and 52.7 ± 25 minutes, respectively.
Most patients (88%) required transfusion perioperatively, with transfused patients receiving a mean of 4.97 ± 3.0 U of packed red blood cells. Fifteen patients (31%) required both fresh frozen plasma and platelets for perioperative bleeding. For patients who were placed on CPB, MAP was maintained at 74 ± 11 mm Hg (range, 52 to 90 mm Hg). Thirty-two patients (65%) required inotropic or vasopressor support during their perioperative course. Mean intensive care unit LOS was 10.8 days (median, 5 days; range 1 to 134 days), and the mean overall LOS was 20.1 days (median, 12 days; range, 4 to 136 days).
In-Hospital Morbidity and Mortality
Arrhythmia was the most common complication, occurring in 57% of patients. One-third experienced respiratory complications, and 25% of these patients required reintubation. Infection occurred in 18% of patients; of these, 50% (n = 5) were leg wounds at the harvest site, 10% (n = 1) were at the sternotomy site, and 40% (n = 4) were other types of infection requiring antibiotic treatment (eg, Clostridium difficile). Other postoperative complications included renal failure, 18%; cerebrovascular accident, 10%; bleeding, 6%; and tamponade, 4%. A permanent pacemaker was required in 16%. At least one postoperative complication occurred in 76% of patients.
The operative mortality, defined as death occurring in-hospital or within 30 days after operation, was 8% (4 of 49). One early death involved a 90-year-old woman who underwent an elective three-vessel CABG and died from cardiac tamponade on postoperative day 3. A second early death occurred in a 90-year-old woman who had an emergency ascending aortic aneurysm repair and two-vessel CABG. The patient experienced perioperative bleeding and ultimately died secondary to a cerebrovascular accident on postoperative day 8. The other two early deaths were a result of multiorgan failure secondary to sepsis. One patient was a 93-year-old woman who underwent an elective MVR and died on postoperative day 12, and the other patient was a 90-year-old man who had an emergency three-vessel CABG and died on postoperative day 132. The in-hospital mortality rate was higher in patients undergoing emergency operations, 13% (2 of 16) compared with patients undergoing elective operations, 6% (2 of 33), but this difference was not significant.
Predictors of Mortality
Single variable Cox proportional hazards models found preoperative chronic renal insufficiency (CRI), EF, severity of heart failure (ie, higher NYHA class), and increasing total cardiovascular comorbidity were associated with increased mortality (Table 2). The forced AIC model selection procedure selected CRI, peripheral vascular disease, and EF as the three best fitting covariates for the multivariate model. As summarized in Table 3, multivariable Cox regression analysis demonstrated CRI (p = 0.007) and decreasing EF (p = 0.031) as independent predictors of death.
|
|
|
Quality of Life
At follow-up in 2005, the 23 patients who were still alive were contacted, and the SF-12 was used to assess quality of life. For these survivors, a mean of 4.0 years had elapsed since their operation (median, 3.6 years; range, 1.2 to 9.5 years). The interview was completed by 12 of the 23 patients (52%), with the remaining survivors unable to complete the interview because of neurocognitive decline in 7, refusal in 3, or language barrier in 1. Because no normative data exist for nonagenarians only, the functional status of our nonagenarian cohort was compared with available normative data for individuals aged 75 years or older, and no statistically significant difference was demonstrated in any of the domains of the SF-12 (Fig 2). When compared with normative data for heart disease patients, our nonagenarian cohort reported a significantly higher score in the General Health domain (p = 0.01) and similar scores in the remaining domains.
|
| Comment |
|---|
|
|
|---|
Patient Selection
Our cohort represents a highly selective subgroup of nonagenarians. Selection criteria were based on our 20 years of experience operating on elderly patients and were not guided by age alone. Previous reports have shared in the observation that it is the physiologic status of the patient, and not the chronologic age, that more closely correlates with clinical outcome [10, 11]. Our preoperative evaluation of patients focused on the level of physiologic and social functioning, the individual patients operative risk, and the patients personal motivation to undergo a surgical procedure.
All patients in our study presented preoperatively with significantly reduced cardiac function (eg, angina, CHF) refractory to medical treatment resulting in an expected 6-month survival of approximately 10%. Physiologic factors considered by the surgeon included baseline liver and renal function, EF, cognitive resilience, previous cardiac surgery, and ambulatory status. Assessment of social functioning included emotional and tangible social support, social integration and community involvement, as well as activity level. Utilization of home health services, such as a nursing home or assisted-living center, by itself did not preclude operation. In general, our exclusion criteria consisted of an operative risk of 50% or more, severe liver or renal disease, an inability to ambulate semi-independently or communicate, and radiographic evidence of significant gray matter loss or recent cerebrovascular accident.
Trends in Perioperative Mortality
Apart from Bridges and colleagues report [28], seven other studies have examined outcomes in nonagenarians undergoing cardiac procedures (Table 4) [10–16]. The earliest of these investigations date back to the 1990s and together demonstrate a mean operative mortality of 12% (range, 7% to 18%) [10–12, 14]. Results of the current study closely reflect the operative mortality of 7% reported in previous work at our center [15] and support the hypothesis that carefully selected nonagenarians can safely undergo cardiac surgical intervention with mortality rates that approach those of younger elderly populations.
|
Postoperative Morbidity: An Ongoing Concern
Despite the promising downward trend in overall mortality, postoperative morbidity remains a significant concern. Our review of existing literature found reported postoperative morbidity among nonagenarians undergoing cardiac procedures of 56% to 100% [10–13, 15, 16].
Recent investigations have concentrated on methods of reducing complications after cardiac operations, including two studies demonstrating the efficacy of prophylactic amiodarone in the reduction of postoperative arrhythmias and arrhythmia-related mortality [29, 30]. In addition, reports have suggested a host of perioperative strategies to curtail the development of postoperative pulmonary dysfunction and nosocomial infection (eg, pneumonia), including optimization of pulmonary dysfunction in at-risk patients [31, 32], use of the off-pump technique in at-risk patients and those with advanced pulmonary disease [33], and use of immunotherapy as an adjunct to antibiotics in the setting of refractory pulmonary infection [34]. Substantial risk reduction of nosocomial infection has also been associated with perioperative decontamination of the nasopharynx and oropharynx [35].
Our current management incorporates several other perioperative strategies to optimize postoperative outcome. In our opinion, the two most important strategies in this population relate to the maintenance of high perfusion pressures while on CPB [36] as well as the fundamental understanding that cardiac procedures in nonagenarians should center on repair of the life-threatening lesion only. To that end, we limit bypass and cross-clamp times. Transfusion requirements are also adjusted to maintain a hematocrit above 30%. Finally, we use an aggressive approach to extubation to promote early ambulation and physical therapy.
Survival Outcomes
Survival outcomes represent an important measure of operative success among nonagenarians. The current study offers a comparison between available data in this population [10–16]. This comparison demonstrates improvement in the field over time relative to similar previous reports.
Mean survival of nonagenarians surviving to discharge has been previously reported to be between 1.6 [16] and 2.9 [10] years, and Blanche and colleagues [11] reported a median survival of 2.6 years. The current study demonstrates continued survival gains, with an overall mean and median survival that exceeds 5 years. Natural life expectancy for our cohort (mean, 91.9 years) is 2.76 years for men and 3.94 for women [37]. Thus, our data demonstrate that nonagenarians can successfully undergo cardiac operations with the potential to surpass or, at the very least, approach their natural life expectancy.
Quality of Life as a Marker of Operative Success
With limited health care resources in a growing managed care environment, cardiac surgery in nonagenarians requires a critical assessment of health-related quality of life measurements. Tsevat and colleagues [38] explored health values of the very old and determined that longevity may not be the primary therapeutic end point in this population. Indeed, there has been a recent movement in the field of medicine toward better appreciation of qualitative outcomes [39, 40].
Quality of life after cardiac operations in nonagenarians has not been well described. In their follow-up on 27 operative survivors, Blanche and colleagues [11] reported 89% were satisfied with their overall perceived functional improvement. Miller and colleagues [12] also noted satisfactory and sustained improvements in all nine of their operative survivors. In the present study, we demonstrate that select nonagenarians are capable of achieving favorable quality of life outcomes similar to that of available normative data for two related populations based on age (ie, 75 years or older) and the disease process (ie, heart disease).
Study Limitations
This study shares in the limitation of previous studies relating to small sample size and the ability to generate multivariate predictors of death from the data therein. Moreover, the heterogeneity of surgical interventions in our cohort precludes similar statistical analyses by operation type. The retrospective nature of this review also serves as a limitation because the absence of a control group prohibits the comparison of alternative treatments during the same time period. In addition, our quality of life investigation included a small, self-selected group of survivors and represents only a cross-sectional assessment in the postoperative setting.
Cardiac Surgery: Should We or Shouldnt We?
Although some clinicians will challenge the benefit of surgical intervention in nonagenarians, cardiac surgery may represent a more efficacious therapeutic modality given that optimal medical treatments and percutaneous coronary intervention (PCI) have failed to reduce the elevated risks in elderly patients. Previous studies have reported a high risk of restenosis for elderly patients undergoing PCI as well as higher mortality rates after PCI than conventional open-heart operations [41–45].
The multifactorial nature of the increased morbidity and mortality observed in nonagenarians likely extends beyond the age-related decline in physiologic function. The suggestion that worse surgical outcomes may be a function of biased selection and subsequent delayed referrals is not without merit [4, 11]. Moreover, delaying elective operations in elderly patients as result of late referral may increase the likelihood of future emergency surgical intervention should invasive measures be later pursued.
Despite significant progress in achieving better survival outcomes, the debate regarding cardiac surgery in nonagenarians will likely persist until such intervention can reliably demonstrate additional improvements in long-term survival, and more important, quality of life. Future efforts should aim to further improve results toward those obtained in younger patients. Longitudinal investigations on the topic of long-term, health-related quality of life in this population are warranted and will likely play an increasingly important role in the evaluation of the efficacy of cardiac operations in elderly patients.
Conclusions
Despite increased perioperative morbidity, carefully selected nonagenarians can undergo cardiac operations and achieve favorable quality of life and survival outcomes that restore their projected life expectancy. Age alone is not an absolute contraindication to surgical intervention; patient selection remains the ultimate predictor of clinical outcome. Accordingly, our results support the potential for more proactive intervention in symptomatic nonagenarians as it relates to earlier consideration of elective, rather than emergency cardiac surgery.
| References |
|---|
|
|
|---|
Related Article
This article has been cited by other articles:
![]() |
D. Wendt, H. Eggebrecht, P. Kahlert, and M. Thielmann Successful transapical aortic valve implantation four weeks before 97th birthday Interactive CardioVascular and Thoracic Surgery, June 1, 2009; 8(6): 684 - 686. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. El Diasty, J. A. Gonzalez, J. Perez, F. Cid, V. Mosquera, J. Cuenca, and A. Juffe Early results of off-pump coronary artery bypass graft surgery using bilateral internal thoracic artery grafts in octogenarian patients during ten years Interactive CardioVascular and Thoracic Surgery, January 1, 2009; 8(1): 104 - 107. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. W. Akins Invited Commentary Ann. Thorac. Surg., March 1, 2008; 85(3): 860 - 861. [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 |