Gender differences in the long-term outcomes after valve replacement surgery.

OBJECTIVE: To compare the long-term outcomes in women and men after valve replacement surgery. DESIGN: Observational study. SETTING: Postoperative aortic valve replacement (AVR) or mitral valve replacement (MVR). PATIENTS: 3118 patients (1261 women, 1857 men) who underwent AVR or MVR between 1976 and 2006 (2255 AVR, 863 MVR), with mean follow-up of 5.6 (4.5) years. MAIN OUTCOME MEASURES: The independent effect of gender on the risk of long-term complications (reoperation, stroke and death) after valve replacement surgery using multivariate actuarial methods. RESULTS: After implantation of an aortic valve bioprosthesis, women had a significantly lower rate of reoperation compared to men (comorbidity-adjusted hazard ratio (HR) 0.4; 95% confidence intervals (CI) 0.2 to 0.9). In contrast, if an aortic mechanical prosthesis had been implanted, women were more at risk for late stroke compared to men (HR 1.7; CI 1.1 to 2.7). After adjustment for age and co-morbidities, women had significantly better long-term survival compared to men after bioprosthetic AVR (HR 0.5; CI 0.3 to 0.6), but there was no survival difference between genders after mechanical AVR. Trends existed towards better survival for women after bioprosthetic MVR (HR 0.6; CI 0.4 to 1.0) and mechanical MVR (HR 0.8; CI 0.5 to 1.1). CONCLUSION: The long-term outcomes after valve replacement surgery differ between women and men. Although women have more late strokes after valve replacement, they undergo fewer reoperations and have better overall long-term survival compared to men.

Assessment of cyclosporine pharmacokinetic parameters to facilitate conversion from C0 to C2 monitoring in heart transplant recipients.

BACKGROUND: Cyclosporine (CsA) 2-hour postdose (C2) monitoring is recommended to assess CsA exposure and predict clinical outcomes among heart transplant recipients. We correlated pharmacokinetic parameters and clinical outcomes in stable long-term heart transplant recipients monitored with C0 to develop an algorithm to convert patients from C0 to C2 monitoring. METHODS: Paired CsA C0-C2 measurements and serum creatinine levels were obtained from 35 heart transplant recipients more than 2 years posttransplantation (mean 8.8+/-4.7 years). RESULTS: The mean CsA dose and C0, C2, and C0/C2 ratio were 85+/-23 mg/12 hours, 123+/-41 ng/mL, 572+/-274 ng/mL and 4.8+/-2.1, respectively. C0 correlated weakly with C2 (r=.42, P=.011). The CsA dose correlated better with C2 (r=.58; P<.001) than with C0 (r=.37; P=.026). A good correlation was noted between C2 and the C2/C0 ratio (r=.73; P<.001), but none between C0 and the C2/C0 ratio. A borderline significant inverse correlation was noted between C0 and the worst endomyocardial biopsy score (r=-.34; P=.045), whereas none was noted with C2. Serum creatinine level did not correlate with either C2 or C0. Among patients with C0 within our target of 100 to 150 ug/L, six had C2 above 300 to 600 ug/L as suggested by the literature. CONCLUSIONS: In long-term heart transplant recipients, we could not identify a single pharmacokinetic parameter that could be used to develop an algorithm to convert from C0 to C2 monitoring; however, C2 may be better than C0 for identifying patients at risk of overexposure to CsA.

Comparative clinical outcomes with St. Jude Medical, Medtronic Hall and CarboMedics mechanical heart valves.

BACKGROUND AND AIM OF THE STUDY: Whether the St. Jude Medical (SJM), Medtronic Hall (MH) or CarboMedics (CM) heart valves confer any relative benefits to patient outcome remains controversial. While numerous studies have analyzed clinical results with a single brand, and a few studies have compared two brands, there are no single-center trials comparing all three valves. METHODS: Our experience with patients who had either a SJM, MH or CM mechanical valve in isolated aortic valve (AVR) or mitral valve (MVR) replacement was reviewed. AVR was performed in 953 patients (SJM = 394, MH = 314, CM = 245) and MVR in 591 patients (SJM = 193, MH = 264, CM = 134). Survivors were assessed annually; follow up consisted of 3336 patient-years (pt-yr) after AVR and 1693 pt-yr after MVR. RESULTS: Preoperatively, in the AVR group, more MH patients had previous valve surgery (p = 0.001) or were in NYHA class III/IV (p = 0.03), and more CM patients had a concomitant surgical procedure (p = 0.005). The hospital mortality after AVR with SJM, MH and CM valves was 3.8, 4.7 and 5.3%, respectively (p = 0.65). In the MVR group, there were more males in the CM group (p = 0.011), more CM patients had concomitant surgery (p = 0.001), and more MH patients had previous surgery (p = 0.006). The hospital mortality after MVR with SJM, MH and CM valves was 8.3, 10.2 and 6.0%, respectively (p = 0.35). There was no late survival advantage in either the AVR or MVR group according to the valve used (p = 0.24 and p = 0.90, respectively). For the AVR group the five-year actuarial freedom from thromboembolism was: SJM 85.8 +/- 2.5%, MH 80.1 +/- 2.7% and CM 85.9 +/- 3.5% (p = 0.04), and for MVR it was: SJM 84.2 +/- 4.0%, MH 77.5 +/- 3.4% and CM 86.9 +/- 5.2% (p = 0.27). Bleeding occurred with a similar frequency in the AVR (p = 0.36) and MVR (p = 0.70) groups. No cases of structural failure were identified in this study. At follow up, among AVR patients NYHA class III/IV was present in: SJM 5%, MH 6% and CM 3% (p = 0.50), while among MVR patients this was identified in: SJM 7%, MH 10% and CM 4% (p = 0.22). CONCLUSION: It is concluded that the SJM, MH and CM mechanical valves offer similar clinical results when used for isolated AVR or MVR. While there is a suggestion of an advantage with bileaflet valves, any differences detected may simply reflect differences in the preoperative patient variables.

The HeartSaver left ventricular assist device: an update.

BACKGROUND: Ventricular assist devices have been shown to be effective as bridges to transplantation and recovery for patients with end-stage heart failure. Current technology has been limited because of the need for percutaneous connections with controllers. The HeartSaver ventricular assist device (VAD) (World Heart Corporation, Ottawa, Ontario, Canada) was developed with the intention of having a completely implantable, portable VAD system. The system consists of an electrohydraulic blood pump, internal and external battery power, and a transcutaneous energy transfer and telemetry unit that allows for power transmission through the skin. Control of the device may be achieved locally or remotely through a variety of communication systems. METHODS: The device has been modified with the Series II preclinical version being available for in vitro (mock loop) and in vivo (bovine model) testing. RESULTS: Seventeen Series II devices have been functional on mock loops or other testing trials for an accumulated 900 days of operation. There have been eight acute experiments using a bovine model to test various components as they have become available from manufacturing. Mean pump output was 10.4 +/- 1.1 L/min in full-fill/full-eject mode. Changes in the last 24 months include (1) cannula redesign for better port alignment and integration of tissue valves; (2) battery redesign to convert to new lithium-ion cells; (3) optimized infrared information and electromagnetic inductance energy transmission through various skin thicknesses and pigmentation; and (4) improved reliability of internal and external controller hardware and software. CONCLUSIONS: Modifications have been required to optimize the HeartSaver VAD's performance. The final HeartSaver VAD design will be produced in the near future to allow for formal in vitro and in vivo testing before clinical implantation.

HeartSaver VAD: a totally implantable ventricular assist device. Results of in vivo studies.

Currently, the most widely utilized ventricular assist devices (VADs) require percutaneous connections and are located either externally (e.g., Thoratec, Abiomed) or intra-abdominally (e.g., Novacor, TCI). These attributes have been implicated in a variety of complications (infection, thromboembolic, gastrointestinal, etc.). To address these concerns, a totally implantable VAD that requires no percutaneous connections and can be implanted in the left hemi-thorax has been developed. The developed device has undergone in vivo evaluation as part of the design and development process. A total of 43 implants in the bovine model, with 5 device versions, have been conducted between July 1992 and February 2000. These studies successfully have demonstrated several important aspects of the developed device, including 1) feasibility of a totally implantable system; 2) capability of the device to support a dysfunctional heart; and 3) ability of the device to provide flows up to 10 L/min in a physiological setting. The studies to date have played a vital role in the design and development process as well as demonstrating the feasibility of a totally implantable intrathoracic VAD. Based on these studies, design optimization was conducted, resulting in the development of the pre-clinical version of the device in preparation for clinical trials.

Determinants of hospital survival following reoperative single valve replacement.

OBJECTIVE: To determine the indicators of risk for hospital death, patients undergoing reoperative valve replacement were analyzed METHODS: Four hundred and eighteen consecutive patients undergoing reoperative valve replacement from 1977 to 1994 were reviewed using univariate and multivariate analysis. RESULTS: Overall hospital mortality was 11.2% with 9.4% mortality with aortic valve replacement and 14.2% with mitral valve replacement (P=0.52). Mortality was 9.7% for patients less than 70 years of age compared with 19.4% for older patients (P=0.03), and was 8.5% for those with anoxia times less than 90 mins versus 21.9% for those with longer anoxia times (P=0.001). For first reoperations, 9.5% of patients died, while for patients undergoing second or more reoperation, mortality was 23.2% (P=0.01). While mortality increased from 8.9% to 19.0% with the addition of a concomitant procedure (P=0.008), it was not affected if the additional procedure was a coronary bypass (P=0. 96). The indication for surgery influenced outcome. Mortality was zero for thromboembolism, 9% for structural failure, 23% for nonstructural failure and 22% for endocarditis (P=0.006). For New York Heart Association (NYHA) functional class I patients, mortality was 1.6% compared with 22.3% for those in NYHA class IV (P=0.006). By multivariate analysis, however, only the indication for surgery and the NYHA functional class influenced survival. CONCLUSIONS: Reoperative valve surgery can be performed with a survival (88.8%) that is similar to the initial procedure (91.2%). The indication for surgery and NYHA functional class alone influenced outcome; therefore, possible early reoperation is indicated before clinical deterioration occurs.

Aortic medial changes associated with bicuspid aortic valve: myth or reality?

OBJECTIVE: To determine whether aortic medial changes are more severe in patients who require aortic valve replacement of congenitally bicuspid aortic valves (BAV) than in patients who require replacement of tricuspid aortic valves and acquired valvular disease (AVD). DESIGN: Aortas from autopsies of 14 patients with BAV and 25 with AVD were histologically assessed by two 'blinded' cardiovascular pathologists and analyzed independently with computer-aided morphometry. The aortic valves were examined for valvular fibrosis and calcification. SETTING: The patient population was from a tertiary-care facility. PATIENTS: Patients were selected by retrospective review of autopsy records for patient deaths after aortic valve replacement, over the period 1984 to 1995. RESULTS: There were no significant differences in age (P=0.89), sex (P=0.94), prevalence of systemic arterial hypertension (P=0.37), valvular degenerative changes (P=0.10 and P=1.0) or heart weights (P=0.60) between the two groups. Histological scores for aortic medial degenerative changes including elastic fragmentation, fibrosis and medionecrosis were not statistically different between the groups. However, morphometry demonstrated less elastic tissue in patients with BAV (P=0.003). CONCLUSION: Routine microscopy shows no significant difference in the degree of aortic medial degenerative changes between patients with BAV and AVD. However, morphometry shows less elastic tissue in the aortas of BAV patients. This may explain the anecdotal increase in aortic fragility and propensity for aortic dissection in these patients.

Circulatory support for cardiogenic shock due to acute myocardial infarction: a Canadian experience.

BACKGROUND: Cardiogenic shock due to acute myocardial infarction (AMI) is associated with high mortality. Circulatory support devices may be used to assist these patients while they await cardiac transplantation. METHODS AND RESULTS: From 1986 to 1997, 25 patients in cardiogenic shock complicating AMI within 3.6+/-0.7 days of the event were supported with artificial hearts. Of the 25 patients, 21 were men with a mean age of 48.4 +/- 1.8 years. The age range was 26 to 62 years. Patients were considered for a device when the following criteria were met: cardiac index less than 1.8 L/min/m2, wedge pressure greater than 20 mmHg despite one or two inotropes and/or intra-aortic balloon support. They received either a CardioWest total artificial heart (n=13), a Thoratec biventricular assist device (n=6) or left ventricular assist device (LVAD) (n=6). Three patients were not considered transplant candidates and died while on the devices (two with multiorgan failure and one found to have a bronchogenic carcinoma after implant), with 22 undergoing cardiac transplantation within 8.6+/-2.2 days of device implant. Six patients died in hospital after the transplants (27.3% mortality). Complications included bleeding or tamponade in seven (28%), pneumonia in six (24%) and right ventricular failure in three LVAD patients (12%). Post-transplant actuarial one-, two- and five-year survival rates were 71.4%, 71.4% and 51%, respectively. CONCLUSIONS: Circulatory support devices offer a means to maintain organ perfusion in patients who develop cardiogenic shock due to AMI. Patients can then undergo transplantation with a reasonable expectation for survival when the alternative is death. Eventually the availability of permanent support devices may obviate the need for transplant in these patients.

Ochronosis: an unusual finding at aortic valve replacement.

The condition known as ochronosis refers to the accumulation of oxidized homogentisic acid in the connective tissues of alkaptonuric patients. The diagnosis is usually made from the triad of degenerative arthritis, ochronotic connective tissue pigmentation and urine that turns dark brown or black on alkalinization. Cardiovascular disease is a less well appreciated aspect of this disorder. A patient with ochronosis of his stenotic aortic valve is reported. The role of the pigment in the genesis of the valve degeneration is discussed.

Progress with the HeartSaver Ventricular Assist Device.

BACKGROUND: Ventricular assist devices (VADs) have been shown to be effective for short- or long-term circulatory support. Devices are either being adapted or newly designed for longer term or permanent support, with the goal to provide patients with improved quality of life. Since 1990, a program has been in place to develop a totally implantable, permanent VAD. METHODS: A multidisciplinary team is developing this VAD with specific goals in mind: (1) that it have an intrathoracic position, (2) that it be a totally implantable device without any percutaneous connections, and (3) that it be possible to communicate with the device from remote locations. These goals would allow for complete patient mobility and flexibility for follow-up. RESULTS: The electrohydraulically actuated VAD combines the blood pump, volume displacement chamber, energy converter, and internal electronic module into a single compact unit. The device called the HeartSaver VAD is powered by a transcutaneous energy transfer system and can be remotely monitored and controlled. Prototypes of different versions of the device have been tested in vitro and in vivo with satisfactory performance. CONCLUSIONS: The prototypes of the HeartSaver VAD have functioned well under test conditions and fulfilled the outlined goals. Further development and testing of the design are being conducted before clinical availability.

Discoordinate modulation of natriuretic peptides during acute cardiac allograft rejection in humans.

BACKGROUND: Increased circulating levels of the cardiac polypeptide hormones atrial natriuretic factor (ANF) and brain natriuretic peptide (BNP) may be observed after orthotopic cardiac transplantation. Both the hypertrophic and inflammatory processes in the allograft may contribute to this increase, but no mechanistic explanation has been suggested for this observation. METHODS AND RESULTS: Plasma immunoreactive ANF and BNP determinations were performed in 10 consecutive transplant patients. These were correlated with degree of rejection as reflected by histopathological findings at serial endomyocardial biopsies. Three patients had associated hemodynamic measurements and blood samples 24 hours before and after transplantation. All rejection episodes that received treatment were accompanied by a marked increase in BNP plasma levels to > approximately 400 pg/mL. Steadily increasing BNP levels preceded overt rejection as assessed by histopathological criteria. The increase in plasma BNP was not always accompanied by an increase in ANF, which suggests the specific upregulation of BNP gene expression during acute rejection episodes. Treatment of the acute rejection episodes led to a substantial decrease of BNP plasma levels. CONCLUSIONS: The significant selective increase in plasma BNP levels found in the present study has not been previously described. This finding provides a new insight into the mechanism of allograft rejection and the modulation of natriuretic peptide synthesis and release. Furthermore, although preliminary, the data suggest that BNP plasma levels could form the basis for a new, noninvasive screening test to predict acute cardiac allograft rejection. Because treatment with the antilymphocyte monoclonal antibody OKT3 (murine monoclonal antibody to the CD3 antigen of the human T-cell) decreased BNP plasma levels, cytokine production by T-cells may mediate the selective increase in circulating BNP.

In vivo evaluation of an intrathoracic ventricular assist device.

In this series of experiments, the Unified System components of the HeartSaver Ventricular Assist Device (VAD) version 5.0 were isolated from the controller and power supply for independent assessment. Five systems with external controller/power supply via a percutaneous lead configuration were tested in 13 male calves (101.8+/-4.3 kg). Two studies were ended acutely because of improper filling and air embolism, respectively. Duration of support was from 2.2 hours to 30 days (mean, 99+/-62 hours). The 30 day survivor was euthanized electively. Study termination was related to postoperative complications in five calves: two with bleeding/tamponade, one with thromboembolism caused by inadequate anticoagulation, and two with respiratory insufficiency. Other causes of termination were: one caused by main building power failure, two from errors in communication between the device and controller, and two caused by hydraulic fluid loss related to housing defects. From these experiments, an intrathoracic position for the calf has been defined, the procedure for implantation without cardiopulmonary bypass has been developed, refinements to the controller have been made, and inflow and outflow cannulae have been reinforced. Hydraulic fluid losses will be solved by proceeding with use of a titanium housing instead of polyurethane. In conclusion, the development of the HeartSaver VAD is progressing, in part because of these experimental and informative animal studies. Further in vivo evaluation of the final version will be conducted before clinical trials.

Development of a ventricular assist device for out-of-hospital use.

BACKGROUND: Success with temporary ventricular assist devices, has prompted interest in devices developed for long term use outside of the hospital setting. METHODS: A totally implantable intrathoracic electro-hydraulic ventricular assist device has been developed. Design focused on providing the recipient with a near normal quality of life. To meet this goal the system utilizes transcutaneous energy transfer and biotelemetry to eliminate percutaneous drive-lines/cables as well as a displacement chamber capable of pressure equalization to atmospheric pressures, so as to eliminate the need for percutaneous venting. An implanted battery provides backup power to allow the recipient the ability to bathe, shower, or swim without connection to an external power source. An integrated telemedicine capability allows the device to be monitored/controlled remotely, using telephone lines. RESULTS: The system has been tested in vitro with early prototypes running for up to 5 1/2 years. The system was studied in calves (n = 25) with durations of support of up to 30 days, demonstrating the ability of the device to function as a totally implantable device without percutaneous connections. CONCLUSIONS: The various in vitro and in vivo studies have demonstrated the feasibility of the totally implantable device. Chronic in vivo experiments will follow in preparation for regulatory submissions for human use.

Are older patients with mechanical heart valves at increased risk?

BACKGROUND: Controversy exists regarding the use of mechanical valves in older patients. Many authorities believe that the use of anticoagulants in the elderly is associated with an increased risk of warfarin-related complications. Therefore, we compared the results with mechanical valves in older patients to a cohort of younger patients. METHODS: Aortic (AVR) or mitral valve replacement (MVR) with a mechanical valve was performed in 1,245 consecutive patients who were followed prospectively. They were grouped by age (group 1, < or = 65 years; group 2, > 65 years). The study groups consisted of AVR (group 1, 459 patients; group 2, 323 patients) MVR (group 1, 313 patients; group 2, 150 patients). RESULTS: The average age for the groups was: AVR (group 1, 51 years; group 2, 70 years; p = 0.03) and MVR (group 1, 53 years; group 2, 70 years; p = 0.03). For AVR the incidence of thromboembolism was 0.050 (group 1) and 0.038 (group 2) (p = 0.37) and the actuarial freedom from thromboembolism was 83.0%+/-3.0% and 86.5%+/-1.0%, respectively (p = 0.13). The incidence of bleeding after AVR was 0.021 for group 1 and 0.028 for group 2 (p = 0.49). For MVR the incidence of thromboembolism was 0.059 for group 1 and 0.051 for group 2 (p = 0.75) and the actuarial freedom from thromboembolism was 78.8%+/-3.0% and 75.4%+/-8.7%, respectively (p = 0.71). The incidence of bleeding after MVR was 0.020 for group 1 and 0.027 for group 2 (p = 0.62). CONCLUSIONS: Mechanical valves perform well in selected older patients with no increased risk of bleeding or thromboembolism.

Delay in revascularization is associated with increased mortality rate in patients with severe left ventricular dysfunction and viable myocardium on fluorine 18-fluorodeoxyglucose positron emission tomography imaging.

BACKGROUND: The identification of high-risk patients who require early revascularization has become increasingly important with the present emphasis on reducing health care resources. This is particularly relevant to health care systems with prolonged waiting times for interventions. Myocardial viability imaging with the use of fluorine 18-fluorodeoxyglucose (FDG) PET may help to identify high-risk patients with severe left ventricular dysfunction. The aim of this study was to evaluate the consequences of prolonged waiting time on cardiac outcomes in patients with left ventricular dysfunction directed to revascularization based on FDG PET imaging. METHODS AND RESULTS: Forty-six patients with coronary disease and an ejection fraction of < or = 35% were considered candidates for revascularization based on FDG PET viability imaging. Thirty-five of 46 patients were subsequently accepted for revascularization. Patients were divided into 2 groups based on the median waiting time after PET: an early group (< 35 days; n = 18) and a late group (> or = 35 days; n = 17). Preoperative mortality rates were significantly increased in the late group (4 of 17 [24%] versus 0 of 18 in the early group; P < 0.05). In postoperative follow-up (17 +/- 7 months), cardiac events occurred in 2 of 18 (11%) and 1 of 13 (7.8%) patients in the early and late groups, respectively. Left ventricular ejection fraction increased after early revascularization (24 +/- 7% to 29 +/- 8%, P < 0.001, baseline versus 3 months) but not in the late group (27 +/- 5% to 28 +/- 6%, P = NS). CONCLUSIONS: Preoperative FDG PET can be used to identify a high-risk group of patients who may benefit from early revascularization. A long waiting time for revascularization is associated with a high mortality rate and suggests that early revascularization is desirable after the identification of hibernating viable myocardium.

Management of patients with mild aortic stenosis undergoing coronary artery bypass grafting.

BACKGROUND: The management of mild aortic stenosis during coronary artery bypass grafting remains controversial. METHODS: We reviewed the medical records of consecutive patients between January 1, 1977, and December 31, 1994, to identify 51 patients with mild aortic stenosis who underwent isolated coronary artery bypass grafting (group A), and 19 patients with mild aortic stenosis who underwent combined coronary artery bypass grafting and aortic valve replacement (group B). Patients with more than moderate aortic regurgitation were excluded. Preoperative angiograms were reviewed to assess the severity of calcification and restricted mobility of the aortic cusps. RESULTS: In group A there were 11 deaths and 8 subsequent aortic valve replacements; in group B there were 5 deaths and 3 prosthetic valve-related complications. There was no difference in event-free survival between the two groups after adjusting for the difference in age. Among group A patients, the initial transvalvular gradient (p = 0.0005) and aortic valvular calcification (p = 0.06) identified patients who demonstrated progression to severe aortic stenosis during follow-up. CONCLUSIONS: Our data suggest that routine aortic valve replacement during coronary artery bypass grafting in patients with mild aortic stenosis is not indicated, but concomitant aortic valve replacement may be appropriate in patients with higher transvalvular gradients and calcified valves.

Cardiac operations in the elderly: who is at risk?

BACKGROUND: With the growing number of elderly patients presenting for cardiac operations we analyzed their early survival data to determine whether any preoperative variables might be indicative of increased risk. METHODS: From 1990 to 1995, 436 consecutive patients who were 75 years old or older had either coronary artery bypass, valve replacement(s), or a combination of these. A total of 34 preoperative variables were assessed for their effect on hospital survival by using univariate and multivariable analysis. RESULTS: There were 266 men and 170 women, with 292 patients being 75 to 80 years old and 144 patients being older than 80 years. Coronary artery bypass was performed in 242 patients, valve replacement was performed in 93 patients, and a combination of these in 101 patients. The operation was considered elective in 202 patients, urgent in 209, and emergent in 25 patients of whom 21 were in cardiogenic shock. Overall there were 61 hospital deaths (13.9%). The most common cause of death, low cardiac output syndrome, occurred in 34 patients of whom 26 suffered a perioperative myocardial infarction. Stroke was the cause of death in eight and multiple organ failure accounted for nine deaths. In the univariate analysis, variables that influenced survival included heart failure (p = 0.004), pulmonary edema (p = 0.004), cardiomegaly (p = 0.02), elevated serum creatinine (p = 0.009), surgical priority (p = 0.002), and cardiogenic shock (p = 0.002). In the multivariable analysis there were three independent determinants of hospital survival: cardiomegaly (odds ratio, 1.8:1) serum creatinine level higher than 150 micromol/L (odds ratio, 5.5:1) and emergency procedure (odds ratio, 2.5:1). CONCLUSIONS: Although cardiac operations can be performed safely in many elderly patients, we identified several factors that might help both in case selection and in perioperative decisions.