YEAR IN CARDIOLOGY SERIES
The Year in Cardiovascular Surgery
Robert H. Jones, MD, FACC*
Department of Surgery/Duke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina.
Manuscript received January 14, 2008;
accepted March 12, 2008.
* Reprint requests and correspondence: Dr. Robert H. Jones, Duke Clinical Research Institute, P. O. Box 17969, Durham, North Carolina 27715. (Email: jones060{at}mc.duke.edu).
The purpose of this fifth and final yearly review of cardiovascular surgery by this author remains to highlight articles that provide new and accurate information that should guide management decisions for patients with cardiovascular disease. In the past, preference has been given to review of reports that link outcomes of patients to decisions about whether an operation should be done. The large body of new evidence about how an operation should be done was considered more appropriate for discussions among cardiac surgeons. However, patients increasingly include questions of how an operation should be done when deliberating about whether they will consent to a proposed operation. This patient interest in decisions about operative conduct now makes topics of how an operation should be performed of broad interest to all cardiovascular specialists. The first 8 of the references in this review focus on questions about how the coronary artery bypass grafting (CABG) operation should be conducted to optimize outcomes for individual patients.
 |
Surgery for Coronary Artery Disease (CAD)
|
|---|
On and off cardiopulmonary bypass (pump) CABG comparison.
The Octopus study conducted at 3 centers in the Netherlands between 1998 and 2000 randomized 281 low-risk patients to receive on-pump CABG (n = 139) or off-pump CABG (n = 142) (1). A battery of 10 standardized and validated neuropsychological tests were administered by a psychologist blinded to treatment 1 day before, at 3 months, 12 months, and 5 years after operation. A cardiovascular event defined as all-cause mortality, stroke, myocardial infarction, or coronary intervention occurred in 30 (21.1%) of off-pump and 25 (18.0%) of on-pump CABG (Fig. 1). The primary outcome defined as cognitive outcome at 5 years could be determined in 123 (94.5%) of 130 off-pump patients and 117 (90%) on-pump patients. No difference was observed between the 2 groups in cognitive decline, angina status, or quality of life.
View larger version (18K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1 Proportion of Patients With a Cardiovascular Event
Cardiovascular events defined as mortality, stroke, myocardial infarction, or repeat coronary revascularization did not differ between off-pump and on-pump coronary artery bypass grafting (CABG) patients. Reprinted, with permission, from van Dijk et al. (1).
|
|
To assess cardiopulmonary bypass as a cause of cognitive decline, cognitive performance at baseline and 3 years was compared in on-pump CABG patients (n = 152), off-pump surgery patients (n = 75), patients with CAD but no surgery (n = 99), and subjects without CAD risk factors (n = 69) (2). Neuropsychological performance was assessed by standardized tests of attention, language, verbal and visual memory, visuospatial, executive function, and psychomotor and motor speed. No group differed significantly in any cognitive domain in comparisons of baseline to 36-month performance. There were no significant differences in the degree of change between on- and off-pump surgery. This study suggests that cardiopulmonary bypass is not an important cause of cognitive decline.
Outcomes of 13,889 off-pump CABG and 35,941 on-pump CABG patients operated between 2001 and 2004 were compared using data prospectively entered into New York State databases (3). Conversion from off- to on-pump CABG was necessary in 226 patients whose operative mortality was 9.7% compared with 2.45% for the total on-pump cohort and 2.30% for the total off-pump cohort. Outcomes of these patients were included in the off-pump cohort. Off-pump patients had less strokes (3.7% vs. 4.2%, p = 0.005) but more unplanned reoperations (1.2% vs. 0.8%, p = 0.047) compared with on-pump CABG patients using risk-adjusted data. Three-year survival was 90.1% for on-pump and 89.4% for off-pump patients (p = 0.2) (Fig. 2). Freedom from subsequent revascularization at 3 years was 93.6% for on-pump and 89.9% for off-pump CABG at 3 years (p < 0.0001) (Fig. 3). Moreover, these outcomes did not change in a subanalysis using 3,074 matched pairs served by surgeons performing at least 50% of their CABG operations by the operative technique their patients represented in the comparison.
View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2 Three-Year Kaplan-Meier Curves for Survival for Off-Pump and On-Pump Bypass Surgery (Matched Cases)
Numbers in the figure are survival rates and 95% confidence intervals. Numbers of patients at risk were 8,867, 6,027, and 3,188 for off-pump coronary artery bypass at 1, 2, and 3 years, respectively, and 8,893, 6,063, and 3,155 for on-pump coronary artery bypass grafting. Reprinted, with permission, from Hannan et al. (3).
|
|
View larger version (13K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3 Three-Year Kaplan-Meier Curves for Freedom From Subsequent Revascularization for Off-Pump and On-Pump Bypass Surgery (Matched Cases)
Numbers in the figure are rates of freedom from subsequent revascularization and 95% confidence intervals. Numbers of patients at risk were 8,466, 5,599, and 2,880 for OPCAB at 1, 2, and 3 years, respectively, and 8,624, 5,742, and 2,938 for on-pump coronary artery bypass grafting. Reprinted, with permission, from Hannan et al. (3).
|
|
The current weight of evidence suggests that off-pump CABG may offer less perioperative morbidity but at the price of a less effective long-term revascularization. However, differences in carefully done randomized and observational studies are sufficiently small that broad generalizations for patient groups become meaningless. Specific baseline characteristics of individual patients will guide selection of the preferable approach for each patient as future data refines criteria for decisions that optimize outcomes for the total CABG cohort treated at each center. The New York State ratio of 2.6:1 on- to off-pump CABG operation appears reasonable with the current state of knowledge.
|
CABG With or Without Mitral Valve Surgery
|
|---|
Prospectively acquired clinical and echocardiographic data in separate databases were merged to identify 390 patients who had severe (3+ or 4+) functional ischemic mitral regurgitation on initial admission during which a CABG was performed with (n = 290) or without (n = 100) a mitral valve annuloplasty (4). In both unadjusted comparisons and 54 propensity-matched patient pairs, survival was similar over 10 years in CABG patients with or without mitral valve annuloplasty (Fig. 4). These well-analyzed data should increase uncertainty among cardiovascular specialists about the need for adding mitral surgery to CABG even in the presence of severe mitral regurgitation. An accompanying editorial suggests the best source of a definitive answer to this question would be an observational study using some large existing database of CABG operations combined with intraoperative telephone notification by the operating surgeon of intent to proceed or not proceed to open the left atrium for a mitral valve repair (5).
View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4 Survival After CABG Either Alone or With Concomitant MV Annuloplasty for Functional Ischemic MR
Symbols represent deaths positioned according to the Kaplan-Meier estimator, vertical bars are 68% confidence limits, and numbers in parentheses are patients still alive. Solid lines are parametric estimates enclosed within 68% confidence limits. (A) Unadjusted survival, based on 37 deaths after coronary artery bypass grafting (CABG) alone and 92 after CABG + mitral valve (MV) annuloplasty. (B) Propensity-matched survival, based on 19 deaths after CABG alone and 19 after CABG + MV annuloplasty. MR = mitral regurgitation. Reprinted, with permission, from Mihaljevic et al. (4).
|
|
|
Single to Multiple Arterial Conduit Comparison
|
|---|
Twenty-year follow-up has been reported on 867 (81%) of the 1,067 consecutive CABG operations performed by 2 cardiac surgeons from 1984 to 1987 (6). During this interval, multivessel CAD was treated with CABG using multiple arterial conduits in 71% of patients by one surgeon and in 26% of patients operated by the other surgeon. Follow-up data were analyzed by treatment received and not by surgeon. The propensity score for operation received was not significant in the Cox model for the end point of death. In the overall group, 71% of patients were dead at 20 years and 52% of these died of cardiac causes. The composite of mortality, myocardial infarction, percutaneous coronary intervention (PCI), and redo CABG was reduced by multiple compared with single internal mammary artery use (p = 0.013).
|
Medication Use in CABG Patients
|
|---|
The risks of 3 hemorrhage-sparing medications commonly used in cardiac surgery were evaluated in a 2-phase observational study (7). Between November 11, 1996, and June 20, 2000, 69 clinical sites enrolled 4,374 CABG patients (18 patients/site) and collected 7,500 fields of baseline and in-hospital data on each patient. Protocol permitted care providers to make decisions for use of hemorrhage-sparing medication. Aprotinin was used in 1,295 (29.6%); aminocaproic acid was used in 883 (20.2%); tranexamic acid was used in 822 (18.8%); and no hemorrhage-sparing medication was used in 1,374 (31.4%) of the 4,374 patients. At close of the in-hospital phase on June 30, 2000, a 4-year program of data collection was begun by 62 of the 69 sites. The 7 centers that did not participate in the follow-up phase had enrolled 498 patients (71 patients/site). Cox regression analysis was performed on all of the 4,374 patients enrolled in the in-hospital phase with the data of surgery as time 0. Patients completing the in-hospital phase but not participating in the follow-up phase were right-censored at the last follow-up time. Multivariable logistic regression with propensity adjustment showed the 87 excess deaths during 5 years of follow-up in the aprotinin-treated group to be significant (p = 0.008), whereas the 26 excess deaths in the aminocaproic acid-treated group and the 8 excess deaths in the tranexamic acid-treated groups were not significant.
The major controversy surrounding these findings focuses on possible confounding from an aprotinin administration decision that itself marks patients who are either more prone to bleed or perceived as less able to withstand the added risks of excessive bleeding should it occur. This concern is reasonable in that aprotinin is the most costly and the most potent of the hemorrhage-sparing agents. Moreover, its use has been associated with excessive intravascular thrombosis in the past. Many experienced cardiac surgeons rarely use any hemorrhage-sparing medication, and few use aprotinin routinely for CABG operations. Centers recruited to participate in this study were requested to enroll only 2% of their expected CABG population. Therefore, the distribution of drug use in this study cannot be considered representative of medication use in the broad CABG population.
A well-conducted meta-analysis compared 8 clinical outcomes of risk and benefit of aprotinin, aminocaproic acid, and tranexamic acid with placebo using data from 138 randomized controlled trials conducted in cardiac surgical patients (8). All agents significantly reduced blood loss by 226 to 348 ml and decreased the proportion of patients transfused. No agent had significant differences in mortality, stroke, myocardial infarction, or renal failure. However, aprotinin at high dose significantly increased the risk of renal dysfunction but also demonstrated significantly less total blood loss than aminocaproic acid (–184 ml) and tranexamic acid (–195 ml). These data provide a more balanced view of the risks and benefits of hemorrhage-sparing drug use in cardiac surgical patients than the previously described observational study.
In the PREVENT (PRoject of Ex-Vivo vein graft ENgineering via Transfection) IV trial, patients free of death or myocardial infarction at 30 days were assessed for medication use after hospital discharge and 1 year and 2 years after CABG (9). Individual patients were categorized as with or without an ideal indication for antiplatelet agents, beta-blockers, angiotensin-converting enzyme inhibitor or angiotensin receptor blockers, and lipid-lowering drugs. Criteria of one-half or fewer, more than one-half but not all, and all indicated medications separated groups of 488 (16.4%), 558 (18.8%), and 1,924 (64.8%) of the 2,970-patient cohort for evaluation of the influence of the quality of secondary prevention medication use on death or myocardial infarction at 2 years. A Cox proportional hazards model adjusting for baseline clinical characteristics and propensity for taking all medications showed a significantly lower death or myocardial infarction rate in patients taking all medications than patients taking one-half or fewer of indicated medications (Table 1).
|
Early Discharge After CABG
|
|---|
The effect of early discharge after CABG on death, readmission, and costs was examined on 55,889 New York CABG patients discharged home from 1995 to 1998 (10). Early discharge was defined as a post-operative stay below the 15th percentile for patients with similar risk. Using this definition based on individual patient risk, 17% of patients were discharged early. This approach normalized outcomes to minimize the tendency of healthier patients at baseline to have less post-operative complications and shorter hospital stays than the total population. The primary clinical outcome of all-cause death (1%) or readmission (19%) within 60 days of discharge after CABG occurred in 20% of patients. About four-fifths of readmissions appeared related to CABG, including infection (20%), heart failure (14%), ischemic events (9%), respiratory illness (7%), arrhythmia (7%), pulmonary embolism or deep vein thrombosis (6%), stroke (4%), and gastrointestinal disorders (2%).
Patients discharged early were less likely to die or be readmitted within 60 days of discharge (15%) than patients with typical (20%) or late discharges (26%). Death rates were 0.44 for early discharge, 0.69 for typical discharge, and 1.71 for late discharge. The percentage of early discharge after CABG among the 32 New York hospitals ranged from 2% to 42%. After adjusting for patient risk, the odds ratio for death or readmission within 60 days ranged from 0.7 to 1.4 (p < 0.001). No association was demonstrated between a hospital's tendency to discharge early and its risk-adjusted in-hospital mortality rate.
In the Medicare subset of this population for whom cost data could be obtained, lower costs of $1,755 within 60 days after early discharge combined with an average savings of $4,554 for the CABG hospitalization resulted in a $6,309 average cumulative savings for patients discharged early relative to those with typical discharges.
These data clearly document that from 1995 to 1998 more patients undergoing CABG in New York State could have safely been discharged home sooner. Now, almost a decade later, early discharge has become more widely accepted as standard practice. Increasingly, the potential medical liability for occurrence of an untoward event soon after discharge is becoming a more dominant concern for cardiac surgeons. Application of methodology used in this report to more modern datasets would better inform optimal timing of discharge decisions and permit patients to leave the hospital for the comfort of home as soon as risk was equivalent in both environments.
|
Comparison of CABG and Percutaneous Transluminal Coronary Angioplasty (PTCA)
|
|---|
The BARI (Bypass Angioplasty Revascularization Investigation) trial.
The 10-year final follow-up of the BARI trial conducted in 1,829 patients showed that 242 (26.5%) of 914 CABG patients died compared with 265 (29%) of 915 PTCA patients (11) (Fig. 5). The statistically insignificant survival advantage of 2.5% for CABG over PTCA added an average of 84 days of survival over the average follow-up of 10.4 years. After further follow-up, the earlier reported survival advantage for CABG over PTCA in the subset of patients with diabetes no longer had statistical validity. Over the 10-year follow-up, the PTCA cohort underwent 3.6-fold more subsequent revascularizations than the CABG cohort.
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5 Overall Survival and Survival Free of Q-Wave MI by Randomized Treatment
The blue solid lines indicate coronary artery bypass grafting (CABG), and percutaneous transluminal coronary balloon angioplasty (PTCA) is indicated by red dashed lines. MI = myocardial infarction. Reprinted, with permission, from the BARI Investigators (11).
|
|
Therefore, the most honest way to communicate BARI results to a patient considering PCI today would be, "If you are among the one-third of patients most suitable for PCI, you may pursue a strategy of PCI to treat your CAD without risk of excess death as long as you accept a likelihood of more subsequent procedures."
Cost effectiveness comparison.
An evidence synthesis approach to cost effectiveness analysis compared PCI stenting with minimally invasive CABG with an internal thoracic artery graft to the left anterior descending coronary artery (12). A meta-analysis of randomized trials comparing the 2 strategies was performed using Cochrane Collaboration methodology. Costs were based on the British National Health Service cost report. A base care analysis for a 61-year-old man was calculated for a 10-year interval in 1-year Markov simulation cycles.
Over 10 years, coronary stenting cost £6,317 and provided 6.72 quality-adjusted life years compared with a cost of £7,146 that provided 6.85 quality-adjusted life years for CABG. This minimal incremental cost effectiveness equates to a cost effectiveness ratio of £6,274 per quality-adjusted life year. Sufficiently small differences demonstrated in this analysis justified the conclusion that cost does not need to be considered in clinical decisions made between these 2 revascularization strategies.
|
Cardiac Surgery for Advanced Heart Failure
|
|---|
Ventricular assist devices (VADs).
A decade of use of VAD implantation for management of shock after cardiac surgery was reviewed using clinical data from the Society of Thoracic Surgeons' National Cardiac Database (13). Between January 1995 and December 2004, 5,735 VADs were implanted in 601 centers. Ventricular assist device implantation occurred in 0.3% of the cardiac operations reported during the decade. Patients who underwent cardiac transplantation during the same hospitalization were excluded to focus analysis on VAD use for failure to wean from cardiopulmonary bypass.
The primary end point, defined as in-hospital death or death within 30 days of hospitalization for cardiac surgery, occurred in 45.9% of patients. Logistic regression identified baseline characteristics known to relate strongly to operative mortality to also predict mortality with VAD use. Baseline characteristics with an odds ratio >2.0 included salvage procedure, reoperation, and dialysis. Age and diabetes were not predictive, and mortality was 56.3% in the 240 patients younger than 40 years and 29.6% in the 311 patients older than 79 years. This observation emphasizes the difficulty in determining the true denominator of those eligible for a high-risk procedure to treat a high-mortality condition. Clearly, the patient group younger than 40 years was more aggressively treated for more hopeless-appearing conditions than the octogenarians who were most likely to be the most salvageable of those in their age group. Fortunately, collection of more detailed information about reasons for VAD implantation began in 2004, so future reports will better inform the appropriate use of VADs in patients who remain in critical condition at the conclusion of a cardiac operation.
A prospective observational study conducted at 7 major U.S. centers enrolled 67 patients who received a left ventricular assist device (LVAD) as a bridge to transplantation and survived at least 30 days between August 2001 and October 2003 (14). The 37 nonischemic and 30 ischemic class IV heart failure patients had a left ventricular ejection fraction (LVEF) of 0.17 ± 0.7. On day 30, the LVEF had increased to 0.34 ± 0.12 but by 120 days had decreased to values observed before LVAD insertion. Decisions to explant LVADs for recovery were individualized at each center. At 134 ± 108 days, 11 (15%) of the 67 patients remained on support devices, 44 (66%) had received transplantation, and 6 (9%) had died. Device explantation for recovery was done in 6 (9%) of patients. All 6 patients remained clinically stable at 6 months. The LVEF remained stable in 4 of these patients.
Between November 2002 and December 2005, data were prospectively entered into a Food and Drug Administration–mandated registry on all patients who underwent HeartMate LVAD (Thoratec Corp., Pleasanton, California) insertion at 66 U.S. hospitals. Data from 280 (91%) of the 309 total patients who gave consent were analyzed to identify predictors of in-hospital mortality (15). The mean age was 60.7 ± 12 years, and 65% of patients had ischemic heart failure. The mean LVEF was 0.176 ± 0.006. Survival estimates were 86.1% at 30 days, 56.0% at 1 year, and 30.9% at 2 years after censoring patients who died (n = 155) or were transplanted (n = 47). A total of 155 (55%) of 280 patients died during a mean support time of 10.4 months (range 1 day to 3.6 years). In-hospital deaths occurred in 76 (64.4%) of 118 deaths within a year of implantation.
A risk score was developed for 90-day mortality using data from 222 patients having complete baseline data. Liver and renal dysfunction, low serum albumin, and hematologic and coagulation abnormalities were the strongest markers of risk. The odds ratios of 9 variables found to be predictive of 90-day in-hospital mortality in multivariable modeling were used to create a simplified risk score (Table 2) that had a C-statistic of 0.89 in the population on which it was derived. The score appeared to aid discrimination of low-, medium-, high-, and very high-risk patients for the end points of 90-day in-hospital mortality and survival to hospital discharge, 90 days, and 1 year (Fig. 6).
View this table:
[in this window]
[in a new window]
|
Table 2 Multivariable Analysis of Risk Factors for 90-Day In-Hospital Mortality After LVAD as Destination Therapy (n = 222)
|
|
Between January 1, 1990, and January 1, 2005, 48,982 patients over the age of 18 years were listed for heart transplantation on the United Network of Organ Sharing (UNOS) waiting list (16). Figure 7
depicts the annual listing of 18,004 UNOS status 1 and 30,978 UNOS status 2 patients and the percentage of listed patients undergoing transplantation each year. Survival on waiting lists significantly improved over this interval. Predictors of deaths by Cox proportional multivariable modeling within 2 months of listing included need for mechanical ventilation, failed heart transplant, status 1A intra-aortic balloon support, and creatinine >1.5 mg/dl for status 1 patients and restrictive cardiomyopathy and intravenous inotropes for status 2 patients. The 1-year survival of status 2 candidates approaches outcomes of heart transplantation, suggesting that early listing of some of these patients might not be justified.
View larger version (31K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7 The Number of UNOS Status 1 and 2 Heart Transplant Candidates Listed in the U.S.: Years 1990 to 2005
The number of United Network of Organ Sharing (UNOS) status 1 candidates listed for heart transplantation in the U.S. decreased from 836 patients in 1990 to 1,159 patients in 2005. During the same time period, the number of patients listed as UNOS status 2 decreased from 2,332 patients listed in 1990 to 1,147 patients listed in 2005. These trends were associated with an increased proportion of the listed-to-transplanted heart transplant candidates per calendar year from 41% in 1990 to 77% in 2005 (red line). Reprinted, with permission, from Lietz et al. (16).
|
|
|
Cardiac Surgery for Congenital Heart Disease
|
|---|
Between November 1, 1994 and October 3, 2005, 164 consecutive pediatric heart transplants were performed on 154 patients with ages ranging from 20 days to 21 years with a median age of 5.2 years (17). Follow-up over 0.4 to 19.5 years averaged 5.9 ± 4.8 years with only 1 patient lost to follow-up. Overall Kaplan-Meier survival was 82.3% at 1 year, 65.3% at 5 years, and 54.4% at 10 years. Dominant causes of the 59 (38.3%) deaths of 164 patients were transplant CAD in 20 (13%), post-operative multisystem organ failure in 15 (9.7%), and rejection in 8 (5.2%). The probability of surviving 1 year was significantly less during the first half compared with the second half of the series. However, survival after 1 year remained constant throughout the entire experience. Cox proportional multivariable models showed need for endotracheal intubation >5 days post-operatively to be the strongest predictor of 1- and 5-year mortality.
Late outcome could be confirmed in 96% of the 6,460 children in the Finnish Research Registry of Pediatric Cardiac Surgery who underwent 7,240 cardiac operations during 1953 to 1989 (18). Death certificates and medical records were used to classify congenital heart disease–related deaths as related to heart failure: sudden, perioperative, and cardiovascular. Each patient's survival was compared with that in the age-, gender-, and time-matched general population each year. On October 29, 1998, the closing date of the study, 5,193 (80%) of the 6,460 operated patients were known to be alive and living in Finland. Death occurred in 1,028 (16%) patients, and outcome was unknown in 239 (4%) patients. Death occurred within 30 days of operation in 836 (7%) patients and more than 30 days in 592 (9%) patients. Figure 8
depicts the cardiac/noncardiac cause of death by diagnosis. A total of 397 (67%) deaths were related and 185 (81%) were unrelated to congenital heart disease. Figure 9
compares survival of the cohort to the matched general population. The 45-year survival was 89% of that of the general population. Male patients had lower survival than female patients. Figure 10
illustrates the mode of cardiac deaths for each diagnosis. The most common mode of death was heart failure usually associated with pulmonary hypertension. Common noncardiac causes of death included neurologic disease and pneumonia usually associated with mental retardation.
View larger version (47K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8 Proportion of Causes of Death
Numbers in blocks define the actual number of deaths in each diagnostic group. ASD = atrial septal defect; COA = coarctation of aorta; Misc = miscellaneous; PDA = patent ductus arteriosus; TGA = transposition of the great arteries; TOF = tetralology of Fallot; UVH = univentricular heart; VSD = ventricular septal defect. Reprinted, with permission, from Nieminen et al. (18).
|
|
View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9 Relative Survival (Survival of Patients/Survival of General Population) of All Patients
Survival of general population is 100%. Dotted lines represent survival for congenital heart defect (CHD)-related causes. Follow-up started from the patient's first operation. Reprinted with permission from Nieminen et al. (18).
|
|
View larger version (45K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10 Proportion of Modes of CHD-Related Death in Different Defect Groups
Perioperative indicates death within 30 days after second, third, or fourth operation. Numbers in blocks define the actual number of death in each group. CHD = congenital heart defect; other abbreviations as in Figure 8. Reprinted, with permission, from Nieminen et al. (18).
|
|
The value of aspirin after operations for complex cyanotic congenital heart disease was evaluated between January 2001 and August 2005 at 15 centers in the U.S., France, and Germany (19). The specific diagnoses of the 1,004 patients enrolled included hypoplastic left ventricular syndrome (n = 346), tricuspid atresia (n = 103), complex tetralogy of Fallot (n = 127), pulmonary atresia (n = 177), heterotaxy syndrome (n = 38), and other forms of complex single ventricle (n = 213). Operations performed included shunts performed without (n = 344) and with (n = 287) cardiopulmonary bypass and Norwood (n = 323) and Sano (n = 50) procedures. Mean age at surgery was 8 days. Aspirin was used in 806 (80%) of 1,002 patients, and its use was not recorded in 2 of the 1,004 infants studied. Outcomes evaluated were time from initial procedure to death, shunt thrombosis, hospitalizations for Glenn or hemi-Fontan surgery at age >4 months, and a composite of the 3 outcomes. Table 3
summarizes these outcomes at 1 year after the initial operation. A Cox regression model showed nonuse of aspirin to be the strongest determinant of death at 1 year, and especially for patients who underwent a shunt on cardiopulmonary bypass or a Norwood procedure (Table 4). Low-dose aspirin (20 mg) and high-dose aspirin (40 mg) were associated with similar mortality rates (Fig. 11).
View this table:
[in this window]
[in a new window]
|
Table 4 Influence of Aspirin Use on 1-Year Mortality of 1,004 Patients Surviving at Least 14 Days After Operation for Cyanotic Congenital Heart Disease
|
|
View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11 Influence of ASA Dose on Survival
Freedom from death, initial procedure to 1 year of follow-up; aspirin (ASA) use for open shunt = shunt on cardiopulmonary bypass, low ASA = 20 mg, high ASA = 40 mg. Reprinted, with permission, from Li et al. (19).
|
|
|
Surgery for Cardiac Valve Disease
|
|---|
Two risk scores for predicting mortality of operations for valvular heart disease were developed and validated using data from the New York State Cardiac Surgery Reporting System (20). The same analytic strategy was used to produce and validate separate risk scores for isolated cardiac valve operations and for cardiac valve operations combined with CABG. The 2001 to 2003 datasets from patients undergoing isolated cardiac valve operation (n = 10,702, in-hospital mortality 4.41%) and cardiac valve operation with CABG (n = 8,823, in-hospital mortality 8.89%) were divided into development and validation cohorts. Risk factors found to be significant predictors of death in the development cohorts were confirmed to be significant in the validation cohorts. Coefficients of these significant variables were normalized to integers. The total possible risk score for any patient ranged from 0 to 49 for isolate valve patients and from 0 to 35 for valve with CABG patients. The accuracy of the 2 risk scores was tested in the 1998 to 2000 datasets from patients undergoing isolated cardiac valve operation (n = 9,662, in-hospital mortality 5.22%) and cardiac valve operation with CABG (n = 8,463, in-hospital mortality 9.32%). After adjusting for mortality differences with the 2001 to 2003 patients, risk scores developed from 2001 to 2003 data performed with acceptable accuracy in predicting mortality and length of stay of the 1998 to 2000 patient cohort.
Table 5
is used to calculate the risk score for an individual patient based on his or her clinical characteristics and the operation to be undertaken. The mortality expected in New York State would be that corresponding to the individual patient's risk score on Table 6
if the operation is to be an isolated valve operation and on Table 7
if the operation is to be a valve with CABG operation.
These risk scores provide a bedside estimate of the mortality any patient would have experienced if they underwent operation at an average New York hospital from 2001 to 2003. These estimates can be adapted for use by any center by adjusting the look-up table mortalities higher or lower by the ratio of their total mortality to the New York total mortality. Moreover, cardiac surgical practices can easily keep running averages of expected to compare with observed mortalities to monitor operative outcomes over time.
The role of valve surgery in left heart infective endocarditis was examined in 546 consecutive patients diagnosed between 1980 and 1998 (21). The primary outcome was all-cause mortality within 6 months of the diagnosis of infective endocarditis. To adjust for survival bias, follow-up times for nonsurgical patients were required to be at least as long as a respective propensity-matched patient. A total of 129 (23.6%) of 546 patients underwent surgery within 30 days of diagnosis. Death occurred in 99 (23.7%) of nonsurgical and 35 (27.1%) of surgical patients. In the subset of patients matched by propensity score, diagnosis decade, and follow-up time, 18 (19.4%) of 93 nonsurgical and 27 (29%) of 93 surgical patients died (p = 0.56). Because the proportional hazard assumption was violated on both time-dependent covariate analyses, a partition analysis was done to distribute the early high risk of death within 7 days of operation. Analysis of the total cohort with the entire risk of the interval between diagnosis and operation assigned to the surgical patients produced a hazard ratio of 1.9 (1.1 to 3.2) (p = 0.02). However, assigning the risk of only the last 3 days of waiting time before operation to the surgical cohort decreased the hazard ratio to 1.5 (0.9 to 2.6) (p = 0.11). Clearly, clinical deterioration caused many of these patients to be taken to surgery in a high-risk state. Future efforts to provide data useful to optimize the use and timing of cardiac surgery in these high-risk patients should define the surgical cohort at the time informed consent documents for operation are signed.
A cohort of 4,617 adult patients who underwent aortic valve replacement by a single surgical group from 1961 to 2003 for aortic stenosis (47%), aortic insufficiency (14%), or both aortic stenosis and insufficiency (39%) were followed through 2005 (22). Concomitant CABG was added to aortic surgery in 34% of these patients, but none had concomitant replacements of other valves. As of 2005, these patients had returned 22,396 yearly questionnaires providing follow-up for up to 41 years. The mean age at operation was 65 (range 20 to 94) years, and 502 patients were octogenarians and 23 patients were nonagenarians.
Among the 4,617 aortic valve replacement patients, 2,382 (51.6%) were dead and 2,235 (48.4%) were still being followed. The complete lifetimes of the living patient cohort were simulated by the Gompertz parametric regression model using age, gender, valve position, and concomitant CABG data from the entire patient group. For the 2,382 patients who died, a total of 17,525 follow-up years were observed (7.4 years mean follow-up). For the 2,235 censored alive patients, a total of 14,146 follow-up years were documented (6.3 years mean follow-up) and an additional 21,652 years of follow-up were imputed using the Gompertz regression model (9.7 years of imputed follow-up). Survival by decade after operation was 50 ± 0.9% at 10 years, 20 ± 1.0% at 20 years, 6.8 ± 0.8% at 30 years, and 3.9 ± 0.8% at 40 years (Fig. 12). Based on reasonable assumptions, the cost effectiveness ratio to aortic valve replacement was $13,528 per quality-adjusted life year and increased with age up to $19,826 for octogenarians and to $27,182 for nonagenarians.
View larger version (18K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12 Influence of Aortic Valve Replacement on Survival
Survival of patients with severe, unoperated aortic valve disease from the literature (Wu et al. [22] references 7 to 9) and survival of 4,617 patients after aortic valve replacement from the authors' series (thick black curve). Reprinted, with permission, from Wu et al. (22).
|
|
A double-blind trial conducted at 3 Finnish University hospitals from August 2005 to June 2006 evaluated 100 mg hydrocortisone and placebo for prevention of atrial fibrillation in 241 consecutive patients undergoing CABG or aortic valve surgery without prior atrial fibrillation or flutter (23). The primary end point of a 5-min or longer episode of atrial fibrillation within the first 84 h after cardiac surgery occurred in 58 (47.9%) of 121 placebo and 36 (30%) of 120 hydrocortisone-treated patients (p = 0.01). Whether hydrocortisone, which presumably acts to decrease the perioperative inflammatory response, or other agents such as beta-blockers that have a more direct myocardial effect, will prove most beneficial to reduce atrial fibrillation in cardiac surgical patients will await further studies.
|
Appendix
|
|---|
For the status of major ongoing randomized clinical trials in cardiac surgical patients, please see the online version of this article.
|
Acknowledgments
|
|---|
The author would like to acknowledge Vanessa Moore for her contributions in the preparation of this manuscript.
|
Footnotes
|
|---|
Dr. Jones is the principal investigator of the National Institutes of Health sponsored STICH (Surgical Treatment for Ischemic Heart Failure) trial.
|
References
|
|---|
1. van Dijk D, Spoor M, Hijman R, et al. Cognitive and cardiac outcomes 5 years after off-pump vs on-pump coronary artery bypass graft surgery JAMA 2007;297:701-708.[Abstract/Free Full Text]2. Selnes OA, Grega MA, Bailey MM, et al. Neurocognitive outcomes 3 years after coronary artery bypass graft surgery: a controlled study Ann Thorac Surg 2007;84:1885-1896.[Abstract/Free Full Text] 3. Hannan EL, Wu C, Smith CR, et al. Off-pump versus on-pump coronary artery bypass graft surgery: differences in short-term outcomes and in long-term mortality and need for subsequent revascularization Circulation 2007;116:1145-1152.[Abstract/Free Full Text] 4. Mihaljevic T, Lam BK, Rajeswaran J, et al. Impact of mitral valve annuloplasty combined with revascularization in patients with functional ischemic mitral regurgitation J Am Coll Cardiol 2007;49:2191-2201.[Abstract/Free Full Text] 5. Jones RH. Adding mitral valve annuloplasty to surgical revascularization does not benefit patients with functional ischemic mitral regurgitation J Am Coll Cardiol 2007;49:2202-2203.[Free Full Text] 6. Rankin JS, Tuttle RH, Wechsler AS, Teichmann TL, Glower DD, Califf RM. Techniques and benefits of multiple internal mammary artery bypass at 20 years of follow-up Ann Thorac Surg 2007;83:1008-1015.[Abstract/Free Full Text] 7. Mangano DT, Miao Y, Vuylsteke A, et al. Mortality associated with aprotinin during 5 years following coronary artery bypass graft surgery JAMA 2007;297:471-479.[Abstract/Free Full Text] 8. Brown JR, Birkmeyer NJO, O'Connor GT. Meta-analysis comparing the effectiveness and adverse outcomes of antifibrinolytic agents in cardiac surgery Circulation 2007;115:2801-2813.[Abstract/Free Full Text] 9. Goyal A, Alexander JH, Hafley GE, et al. Outcomes associated with the use of secondary prevention medications after coronary artery bypass graft surgery Ann Thorac Surg 2007;83:993-1001.[Abstract/Free Full Text] 10. Cowper PA, DeLong ER, Hannan EL, et al. Is early too early?. Effect of shorter stays after bypass surgery. Ann Thorac Surg 2007;83:100-107.[Abstract/Free Full Text] 11. The BARI Investigators The final 10-year follow-up. Results from the BARI randomized trial. J Am Coll Cardiol 2007;49:1600-1606.[Abstract/Free Full Text] 12. Rao C, Aziz O, Pansesar SS, et al. Cost effectiveness analysis of minimally invasive internal thoracic artery bypass versus percutaneous revascularization for isolated lesions of the left anterior descending artery BMJ 2007;334:621-627.[Abstract/Free Full Text] 13. Hernandez AF, Grab JD, Gammie JS, et al. A decade of short-term outcomes in post cardiac surgery ventricular assist device implantation: data from the Society of Thoracic Surgeons' National Cardiac Database Circulation 2007;116:606-612.[Abstract/Free Full Text] 14. Maybaum S, Mancini D, Xydas S, et al. Cardiac improvement during mechanical circulatory support: a prospective multicenter study of the LVAD working group Circulation 2007;115:2497-2505.[Abstract/Free Full Text] 15. Lietz K, Long JW, Kfoury AG, et al. Outcomes of left ventricular assist device implantation as destination therapy in the post-REMATCH era: implications for patient selection Circulation 2007;116:497-505.[Abstract/Free Full Text] 16. Lietz K, Miller LW. Improved survival of patients with end-stage heart failure listed for heart transplantation J Am Coll Cardiol 2007;50:1282-1290.[Abstract/Free Full Text] 17. Morales DLS, Dreyer WJ, Denfield SW, et al. Over two decades of pediatric heart transplantation: how has survival changed? J Thorac Cardiovasc Surg 2007;133:632-639.[Abstract/Free Full Text] 18. Nieminen HP, Jokinen EV, Sairanen HI. Causes of late deaths after pediatric cardiac surgery. A population-based study. J Am Coll Cardiol 2007;50:1263-1271.[Abstract/Free Full Text] 19. Li JS, Yow E, Berezny KY, et al. Clinical outcomes of palliative surgery including a systemic-to-pulmonary artery shunt in infants with cyanotic congenital heart disease: does aspirin make a difference? Circulation 2007;116:293-297.[Abstract/Free Full Text] 20. Hannan EL, Wu C, Bennett EV, et al. Risk index for predicting in-hospital mortality for cardiac valve surgery Ann Thorac Surg 2007;83:921-930.[Abstract/Free Full Text] 21. Tleyjeh IM, Ghomrawi HMK, Steckelberg JM, et al. The impact of valve surgery on 6-month mortality in left-sided infective endocarditis Circulation 2007;115:1721-1728.[Abstract/Free Full Text] 22. Wu Y, Jin R, Gao G, Grunkemeier GL, Starr A. Cost-effectiveness of aortic valve replacement in the elderly: an introductory study J Thorac Cardiovasc Surg 2007;133:608-613.[Abstract/Free Full Text] 23. Halonen J, Halonen P, Jarvinen O, et al. Corticosteroids for the prevention of atrial fibrillation after cardiac surgery. A randomized controlled trial. JAMA 2007;297:1562-1567.[Abstract/Free Full Text]
|
Top
Surgery for Coronary Artery...
CABG With or Without...