| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Thoracic Surgery, Appalachian Regional Healthcare System, South Williamson, Kentucky
b Louisiana State University School of Medicine, New Orleans, Louisiana
c Department of Cardiothoracic Surgery, The University of Tennessee HSC, Memphis, Tennessee
d Department of Cardiothoracic Surgery, Baptist Memorial Hospital, Memphis, Tennessee
* Address correspondence to Dr Garrett, Cardiovascular Surgery Clinic Inc, 6029 Walnut Grove Rd, Suite 401, Memphis, TN 38120 (Email: egarrettmd{at}cvsclinic.com).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsSTS Database PublicationsSTS Database HistorySTS Database Scientific...CommentReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsSTS Database PublicationsSTS Database HistorySTS Database Scientific...CommentReferences |
|---|
The STS database was initially created to collect patient data regarding coronary artery bypass graft (CABG), surgery, and the majority of scientific work conducted that relates to this procedure. This study reviews the genesis, history, and growth of the STS database, and an organized review is presented of its scientific contributions, primarily in coronary revascularization.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsSTS Database PublicationsSTS Database HistorySTS Database Scientific...CommentReferences |
|---|
|
STS Database Publications |
|---|
|
TopAbstractIntroductionMaterial and MethodsSTS Database PublicationsSTS Database HistorySTS Database Scientific...CommentReferences |
|---|
The breakdown of scientific publications were as follows: adult cardiac surgery (52), general thoracic surgery (3), and congenital cardiac surgery (4). In the adult cardiac surgery category, the subject of publications included CABG (31), valve surgery (12), transmyocardial revascularization (2), atrial fibrillation surgery (1), ventricular assist devices (1), and surgical ventricular restoration (1). The remaining studies (4) involved patients from mixed surgical groups.
|
STS Database History |
|---|
|
TopAbstractIntroductionMaterial and MethodsSTS Database PublicationsSTS Database HistorySTS Database Scientific...CommentReferences |
|---|
The STS created the National Database Audit and Validation Subcommittee to ensure accuracy and completeness of the data submitted, and to establish a database training institute in Minneapolis, intended to be used to train data managers from participating surgical programs. This subcommittee established minimum thresholds of completeness for core data elements prior to accepting the groups' data into the aggregate database. Consequently, the software vendor included internal data quality reporting features to assure completeness and a national nurse coordinator was hired to communicate directly with local data managers. In 1997, a data manager training session was included in the STS annual scientific meeting [4].
From 1990 to 1997, the STS database was maintained by Summit Medical, but in 1997, an STS executive decision licensed multiple software vendors to better address the needs of participating institutions. In 1998, the data warehouse functions transferred from Summit Medical to the Duke Clinical Research Institute to be associated with a reputable academic institution for data analysis and risk stratification. The data was transmitted electronically contingent on front-end software data checks and back-end checks by the institute. Presently, the harvest of data is analyzed and benchmarked with national and regional aggregate data, and site-specific reports are sent to participating institutions semi-annually [5].
In 1997, the STS created the Definitions Subcommittee under the leadership of Dr Thomas Bruce Ferguson, Jr. In concert with the American College of Cardiology, Ferguson, Jr, revised the definitions for both society databases creating new core and extended datasets. The core dataset is mandatory for accurate clinical practice representation, analysis, and risk model development. The extended dataset collects nonmandatory and less critical, but still important data [5].
In 1991, Dr Richard Clark [6] published the first STS database report announcing the successful commencement of this endeavor. Summit Medical contributed 57,555 records from a previously owned database covering patients from 1980 through 1990; thereafter, prospective data was entered and significant growth followed from 1991 to 1993. The first annual report to subscribers containing data from 1980 through 1991 was released in January 1992 [1, 7].
The thoracic surgical community is indebted to Clark [6] as the pioneer who directed the development of the STS database. For 2 decades, it has grown as a versatile tool to benchmark outcomes for individual surgeons and provide large scale scientific contributions [8].
|
STS Database Scientific Publications |
|---|
|
TopAbstractIntroductionMaterial and MethodsSTS Database PublicationsSTS Database HistorySTS Database Scientific...CommentReferences |
|---|
In December 1994, Clark [3] reviewed the preoperative characteristics of the STS database population of a decade and reported a divergence of the expected and observed mortality rates; these findings warranted the creation of a new risk stratification model with annual updates expected to occur thereafter [7].
Risk Stratification Models
The STS database has published the following seven CABG-only risk stratification models: in 1994, covering 1984 to 1990 [9]; in 1997, covering 1990 to 1994 [11]; in 1998, covering 1995 [12]; in 1999, covering 1996 [13]; and in 2002, covering 1990 to 1999 [14]. These risk models determined the expected risk of procedural mortality. In 2002, covering 1997 to 2000, the risk model calculated the expected length of hospital stay and risk of short or prolonged hospitalization [15]. In 2003, covering 1997 to 1999, in addition to mortality risk, the model calculated risk-adjusted rates for five major morbidity endpoints: permanent stroke, renal dysfunction or failure requiring dialysis, any cardiac surgery reoperation, prolonged ventilation greater than 48 hours, and deep sternal wound infection [16].
The first risk stratification model in valve surgery was published in 1999 covering valve replacement patients with or without CABG from 1986 to 1995. This model used 51 preoperative variables to calculate operative mortality for isolated aortic valve replacement, isolated mitral valve replacement, multiple valve replacement, aortic valve replacement with CABG, mitral valve replacement with CABG, multiple valve replacement with CABG, and valve replacement with other thoracic procedures. Logistic regression analysis identified 30 independent variables associated with increased mortality and odds-ratios (OR) ranging from 1 to 7. Salvage status, renal failure, emergent status, multiple reoperations, and New York Heart Association functional class IV were associated with the highest OR for operative mortality. Furthermore, concomitant CABG doubled the mortality rates for all age groups and valve models, and older age increased mortality significantly for isolated aortic valve replacement in patients older than 70 years, and for isolated mitral valve or multiple valve replacement in patients older than 60 years. Although prior single institution studies have reported endocarditis as a significant risk factor, this variable was not included in the risk model design [17].
Surgical Volume as an Indicator of Outcomes
In 1996, Clark [18] reviewed CABG-only operations from 1991 to 1993, finding no correlation between surgical volumes and operative mortality except in the smallest volume groups (<100 cases/year). In addition, lower volume practices (<600 cases/year) had an excessive variability of mortality rates, whereas the higher volume groups (>600 cases/year) showed a narrow range of mortality. Although it is statistically valid that the smallest volume group (<100 cases/ year) had less favorable outcomes, the significant variability in mortality among lower volume practices (<600 cases/year) reflects an unreliable volume-outcome relationship. The expected mortality did not differ according to the size of the practice supporting the notion that case-mix is not a function of patient volume.
In 2004, Peterson and colleagues [19] concluded that operative mortality was only modestly improved by CABG volume, without correlation in patients younger than 65 years or at a low operative risk (0.07% for every 100 additional patients per year). The high variability in mortality among hospitals with lower volume prevented the establishment of a reliable volume threshold to discriminate high-risk centers versus low risk centers. To expand Clark's findings, closing low-volume centers (<150 cases/year) would only prevent 45 CABG procedural deaths annually. Mortality rates were comparable in patients reported to both the STS and Centers for Medicare and Medicaid Services databases; however, when comparing STS participants and nonparticipants from Centers for Medicare and Medicaid Services data, mortality rates favored the STS-participant sites (4.5% vs 5.2%).
In 2006, Disesa and colleagues [20] studied the impact of state certificate-of-need regulations for cardiac surgery pertaining to patient volumes and outcomes after CABG. Institutions from certificate-of-need states recorded higher CABG volumes, but after adjusting for patient risk factors, region, random state effects, and population density, operative mortality was not significantly different. STS participants seemed to have higher patient volumes. To prevent selection bias, the analysis was repeated in the Centers for Medicare and Medicaid Services dataset, finding no mortality difference with respect to certificate-of-need status. These findings question the value of certificate-of-need regulations as a measure of quality improvement, and (as other STS-based studies have concluded) do not support the contention that volume is an indicator of quality.
Gender Relationship to Outcomes
In 1998, Edwards and collegues [21] investigated the relationship of gender to outcomes in CABG-only procedures, using the 1994 risk model to stratify patients into seven risk categories. Gender was removed during model derivation to prevent confounding during the stratification process. Both genders were compared in equally risk-matched categories and women had a significantly higher mortality throughout the risk spectrum, except in patients with an expected operative mortality greater than 30%. This analysis was repeated in patients receiving an ITA graft to exclude the bias of ITA underuse in women; the gender-related mortality difference remained present despite this adjustment. Interestingly, there was an association between body surface area and gender-related mortality, with the mortality difference between men and women more pronounced at higher body surface area levels and disappearing in the lower body surface area groups.
In 2001, Hogue and colleagues [22] studied the effect of gender in neurologic events (ie, stroke, transient ischemic attack, or coma, or a combination of these) after CABG or valve surgery. After adjusting for preoperative risk factors, female gender was associated with an increased rate of neurologic events (odds ratio [OR], 1.21). Neurologic complications carried a seven-fold increased mortality risk with a persistent higher mortality in women (33% vs 28%). Interactions between gender and other variables classically associated with postoperative neurologic events were also examined but were not found to be significant [22].
In 2003, Haan and colleagues [23] studied the impact of gender in the elderly population (>75 years) after CABG-only operations. A multivariate logistic regression model excluding gender was used to stratify patients into five quintiles of increasing mortality risk; operative mortality was compared between genders in each risk category. There was an increased mortality in women throughout the risk spectrum, but gender association with operative morbidity was less consistent.
Race Relationship to Outcomes
The original STS risk models did not identify an association of race with operative mortality. This association was first reported in the 1996 STS CABG-only risk model after race was re-classified into Caucasian or non-Caucasian because of limited records within other racial subgroups [13].
In 2000, Bridges and colleagues [24] studied the effect of race on operative mortality after isolated CABG. A multivariate logistic regression model excluded race stratified patients into risk deciles. Both race subtypes were compared throughout the risk spectrum and the black race was associated with an increased risk-adjusted mortality (OR, 1.29); however, this difference was greatest in the lowest risk decile (OR, 1.83) and progressively faded toward the highest risk decile (OR, 1.03). Black patients were more likely younger, female, hypertensive, diabetic, and in heart failure; therefore, separate risk models were created for each race category, and the magnitude and directionality of risk factors were similar in both subsets. Further scrutiny for risk factor interactions revealed race-by-gender as significant, indicating that the race-related mortality difference is more pronounced in men; in fact, operative mortality showed no significant difference between races in women.
In 2001, Hartz and colleagues [25] published a similar study on the effect of race and the race-by-gender interaction after isolated CABG. As before, patients were divided into Caucasians and non-Caucasians reaching similar conclusions; operative mortality was higher in non-Caucasians and in women. There was also a correlation of predicted operative mortality risk for both race and gender; however, a predicted risk higher than 10% for race and 5% for gender eliminated these mortality differences.
Off-Pump CABG
In 2001, Cleveland and colleagues [26] compared risk-adjusted outcomes after off-pump and conventional on-pump CABG by creating multivariate logistic regression models for operative mortality and the five established morbidities defined by the STS. The risk-adjusted mortality and major complication rates in the conventional group were 2.93% and 14.15%, and in the off-pump group were 2.31% and 10.62%, respectively. The observed-to-expected ratio of operative mortality was 0.81 in the off-pump group and 1.02 in the conventional group (p < 0.05), equivalent to a 19% reduction. The observed-to-expected ratios for each major complication also favored the off-pump group protecting against all endpoints, except for deep sternal wound infection. Furthermore, three patient populations were evaluated; pre-existing renal insufficiency (creatinine >1.5), known cerebrovascular disease, and preoperative chronic obstructive pulmonary disease; the endpoints of postoperative renal failure, cerebrovascular accident, or coma, and prolonged mechanical ventilation were calculated, respectively. The rates of postoperative stroke and prolonged mechanical ventilation were significantly lower in the off-pump group compared with the conventional group.
In 2007, Puskas and colleagues [27] compared the gender-related disparity in outcomes after off-pump and conventional CABG. Whereas, Cleveland and colleagues [26] used 1998 to 1999 data with 9.9% of patients in the off-pump group, Puskas and colleagues [27] used 2004 to 2005 data with more than 20% of patients in the off-pump category. Furthermore, the STS added a field for intraoperative conversions by 2004, allowing an intention-to-treat design by including these conversions in the off-pump group. Off-pump CABG resulted in a statistically significant reduction in risk-adjusted mortality and postoperative complications in women, whereas in men, the difference only reached statistical significance for postoperative complications. This publication concludes that off-pump CABG decreases the gender disparity in outcomes [27].
STS Database Versus Administrative and Mandatory Databases
In 2001, Grover and colleagues [5] compared the Department of Veterans Affairs and the STS national databases, and found striking similarities in the predictive risk factor profiles and ORs. The observed-to-expected ratios also decreased spanning a decade, despite worsening risk factor profiles in both databases. The Veterans Affairs database is mandatory, 99% male, and is underwritten by the federal government. The STS database is a voluntary effort, 77% male, and is financially independent of state or federal funds; therefore, unlike the Veterans Affairs database, the Freedom of Information Act can not mandate the release of STS data to the public, protecting its confidentiality [5].
In 2007, Welke and colleagues [28] compared surgical volumes and mortality rates between the STS and the Medicare administrative claims databases. Patients older than 65 years who were in hospitals common to both databases were compared for CABG, aortic valve replacement, and mitral valve replacement. Comorbidities were coded more frequently in the STS database reflecting differences in data definitions, abstraction methods, and quality data acquisition. Mortality rates reported in both databases were comparable, and for STS participants, the surgical volumes were generally higher, and mortality rates were lower compared with nonparticipant groups. These findings suggest that underreporting is unlikely in the STS database, and this proves its suitability for tracking national cardiac surgery outcomes [28].
Obesity
In 2002, Prabhakar and colleagues [29] studied the influence of moderate (body mass index [BMI], 35 to 39.9) and extreme obesity (BMI >40) after CABG. Preoperative risk was adjusted by multivariate logistic regression after excluding body surface area from the latest STS risk model. Operative mortality, reoperation, deep sternal wound infection, renal failure, and prolonged postoperative hospital stay increased modestly for moderate obesity (OR, 1.21) and significantly for extreme obesity (OR, 1.58). Interestingly, cardiopulmonary bypass and cross-clamp times were similar across the BMI spectrum. There was a "U-shaped" continuous relationship between BMI and risk-adjusted mortality in four categories: men, women, younger (<75 years), and older patients (>75 years); thus, the operative risk increases at the lowest and the highest BMI range. This was the first study to reveal poor outcomes associated with an elevated BMI due to the limited samples collected in the high BMI range in previous studies.
Diabetes
In 2002, Carson and colleagues [30] found that patients with diabetes had an increased risk-adjusted operative mortality after CABG; more pronounced in the insulin dependent group (OR, 1.39), and less, but still statistically higher, in the oral hypoglycemic group (OR, 1.13). Major complications that were analyzed included postoperative myocardial infarction, renal failure, infection, stroke, and multisystem organ failure. Risk-adjusted postoperative morbidity mirrored the pattern found with operative mortality and was approximately 35% higher in diabetics than nondiabetics, and higher in the insulin-dependent compared with the oral hypoglycemic group.
In 2007, Savage and colleagues [31] compared the use of one versus both ITA grafts on deep sternal wound infection in diabetics by using a multivariate logistic regression model to adjust for preoperative risk. There was a statistically significant increment in deep sternal wound infection in diabetic versus nondiabetic patients, further increased by using both versus one ITA graft (2.8 % vs 1.7 %; OR, 2.16), without a significant mortality difference [31].
Process Measures in CABG
In 2002, Ferguson reported on the use of preoperative B-blockers and their effect on operative morbidity and mortality following CABG. Major predictors of B-blocker use were determined using a random-effects logistic model. The influence of B-blockers on morbidity and mortality was measured by using direct risk-adjustment with a multivariable logistic regression model, and by a matched-pairs analysis based on propensity for preoperative B-blocker therapy. B-blocker use was associated with slightly lower mortality (OR, 0.94), less prolonged ventilation (OR, 0.95), and decreased renal failure (OR, 0.91) after adjusting for patient risk factors and site effects (hospital provider). Among patients with an ejection fraction less than 30 %, however, preoperative B-blocker use showed a trend towards higher mortality rates (OR, 1.13; p = 0.23). Furthermore, risk-adjusted mortality declined as B-blocker use increased across sites, thus, suggesting preoperative B-blockers as a useful process measure in quality improvement [32].
In 2002, Ferguson and colleagues [33] studied outcomes associated with ITA grafting in patients 75 years or older using multivariable logistic regression and matched-pair propensity scores adjusted for confounding patient and site-related variables. The ITA grafting significantly decreased both mortality (OR, 0.85) and prolonged ventilation (OR, 0.86). Although the incidence was low in both groups (0.52% vs 0.66%), a significant increase in deep sternal would infection occurred (OR, 1.36). The ITA grafting showed an operative mortality benefit throughout the preoperative risk spectrum, but it seemed to decline in patients older than 85 years or patients with an ejection fraction less than 45% (p > 0.5).
Age in CABG
In 2003, Bridges and colleagues [34] reported on cardiac surgery in nonagenarians and centenarians using data from 1995 to 2000. There were 5 patients 100 years or older, 1,092 patients 90 to 99 years, and 59,576 patients 80 to 89 years who underwent CABG or valve surgery, or both. The 5 centenarian patients underwent isolated CABG and survived without complications, except for prolonged ventilation in 1 patient. The mortality rates in the groups of 90 to 99 years and 80 to 89 years were 11.8% and 7.1%, respectively. Multivariate logistic regression was used to develop a risk model in the CABG-only nonagenarian patients. The predictive factors with the highest contribution to mortality were emergent or salvage circumstances, preoperative intra-aortic balloon pump, renal failure, peripheral vascular disease, and cerebrovascular disease. Patients lacking these risk factors accounted for 57% of the group and had a mortality rate of 7.2%. Female gender was not associated with increased mortality.
Renal Function in CABG
In 2006, Cooper and colleagues [35] studied the impact of renal dysfunction on CABG outcomes by conducting a multivariate logistic regression analysis with glomerular filtration rate as the study variable. Patients were stratified into mild, moderate, severe renal dysfunction, and dialysis dependence. Risk-adjusted operative mortality rose inversely with declining renal function, from 1.3% in mild to 9.3% in severe dysfunction. The operative mortality risk and continuous glomerular filtration rate followed a two-slope nonlinear relationship with an inflection point of approximately 60 mL/min/1.73 m2. A similar risk-adjusted pattern was observed with the STS established postoperative morbidities, with an increased risk that paralleled a declining glomerular filtration rate. Whereas prior studies have considered renal function as a dichotomous variable, the large size and clinically-rich information of the STS database permitted the use of glomerular filtration rate as a continuous variable [35].
In 2006, Mehta and colleagues [36] developed a bedside tool to estimate the probability of postoperative dialysis in patients undergoing CABG or valve surgery, or both, excluding patients on dialysis prior to the procedure. Logistic regression identified major predictors for dialysis, and their regression coefficients were entered into a simplified 10-variable additive risk score model. Postoperative dialysis ranged from 1.1% for CABG to 5.1% for CABG plus mitral valve surgery. There was wide overlap in preoperative creatinine among patients requiring dialysis compared with those not requiring dialysis (median, 1.8 vs 1.1 mg/dL); however, mortality was markedly dissimilar (45% vs 2.5%, respectively). Although preoperative creatinine was higher in patients requiring dialysis, it could not discriminate the presence of this complication, because a large proportion of patients with elevated creatinine did not require postoperative dialysis.
Emergency CABG
In 2006, Haan and colleagues [37] studied the outcome for a decade (1994 to 2003) of emergency CABG within 6 hours of percutaneous coronary intervention. There was a decline in emergency CABG from 2.9% to 0.8% of all CABG patients during this timeframe with an increase in the risk profile and risk-adjusted morbidity and mortality, especially in the post-infarction patient whose volume remained constant at 30% of the emergency CABG volume. In spite of the critical condition of these patients, there was an increase in ITA grafting and the number of distal anastomoses performed. Although there was a steady increase in the volume of percutaneous coronary intervention, the rate of emergency CABG continued to decline.
In 2008, Mehta and colleagues [38] created a simplified bedside additive risk score using 11 preoperative variables to estimate the operative mortality for patients in cardiogenic shock undergoing CABG. Logistic regression calculated weights for the risk score variables, and the final model retained a strong predictive power for mortality. The proportion of patients presenting in cardiogenic shock was 2.1%, and their operative mortality was 22%. This incidence remained stable during the 4-year study period. Although the patients presented in critical condition, ITA grafting was still used in 57%. Although age was strongly associated with mortality, 70% of patients survived this surgery in the oldest group (>75 years). This study underscores the ability to accrue a substantial patient volume in an infrequent clinical condition and attests to the robust and versatile scientific potential of the STS database.
Valve Surgery
The STS has reported 12 scientific studies in valve surgery. The first risk stratification model was published in 1999 covering data from 1986 to 1995. In 2001, Edwards updated the model by using a contemporary and more homogeneous population from 1994 to 1997 [17, 39]. In 2002, Mehta and colleagues [40] analyzed the influence of age on outcomes after mitral valve replacement, creating a classification tree to identify low-risk elderly patients. In 2003, Savage reported a 10-year STS experience of the frequency and patterns of mitral valve repair [41]. In 2004, Haan and colleagues [42] studied the impact of a low ejection fraction (<30%) on outcomes in mitral valve replacement and proposed a classification tree to identify a low-risk patient subgroup. In 2005, Taylor and colleagues [43] studied the influence of race on morbidity and mortality after valve replacement; Gammie and colleagues [44] published the STS experience with mitral valve endocarditis; and Rankin and colleagues [45] presented a comprehensive multivariable logistic regression model of operative mortality in valve surgery. In 2007, Bridges and colleagues [46] studied the association between prosthesis internal orifice size and operative mortality in aortic valve replacement, concluding a lack of correlation between these variables [46]. In 2007, Gammie and colleagues [47] studied the influence of procedural volume in valve surgery for mitral regurgitation on the rates of mitral valve repair, bioprosthetic use, and operative mortality. In 2008, Song and colleagues [48] reported on the age-related mortality difference between men and women after mitral valve surgery, and found an increased risk-adjusted mortality in women aged 40 to 60 compared with men; however, this survival disparity declined with further aging. In 2009, Brown and colleagues [49] reported a 10-year STS experience in isolated aortic valve replacement and determined trends in patient risk profile, risk factors, and surgical outcomes.
Miscellaneous
In 2007, O'Brien and colleagues [50] tested the Centers for Medicare and Medicaid Services Hospital Quality Incentive Demonstration Project, which is a score system proposed to apply financial rewards or penalties based on performance, using the STS database. This composite quality score consists of process measure data (ie, use of B-blockers, ITA, and so forth, and risk-adjusted mortality data). There was substantial disagreement among hospital rankings after testing several alternative performance score systems; consequently, depending on the score methodology used, there was significant variation in hospitals rewarded or penalized on a pay-for-performance scale. Thus, it is of utmost importance to create a reliable composite measure score and validate it by empirical testing.
In 2007, Haan and colleagues [51] studied the impact of residency status on outcomes after isolated CABG, and reported perfusion and cross-clamp times marginally higher in residency versus nonresidency programs. In addition, longer perfusion times were collectively associated with higher mortality rates. In spite of these findings, the risk-adjusted mortality was not significantly different according to residency status (OR, 0.96; p = 0.51); in fact, operative mortality in most perfusion time categories was marginally lower in residency programs. Furthermore, both programs had relatively stable perfusion times during the study period; thus, the decreasing operative mortality over the reported time frame may therefore be due to other unmeasured factors [51].
In 2005, Fowler and colleagues [52] created a bedside scoring system to estimate the risk of major infections after CABG (ie, mediastinitis, vein harvest site infection, or septicemia). Unlike previous STS-based models, Fowler and colleagues [52] developed two scoring systems, one based on preoperative data (12 variables), and another combining preoperative and intraoperative data (12 + 3 variables). In the preoperative model, BMI >40 had the strongest predictive value, followed by cardiogenic shock, renal failure, and concomitant surgery. In contrast, in the combined scoring system, BMI >40, perfusion time >200 minutes, and intra-aortic balloon pump use had the highest predictive values [52].
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsSTS Database PublicationsSTS Database HistorySTS Database Scientific...CommentReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsSTS Database PublicationsSTS Database HistorySTS Database Scientific...CommentReferences |
|---|
75 years of age undergoing coronary artery bypass grafting Am J Cardiol 2003;91:1255-1258.[Medline]Related Article
This article has been cited by other articles:
|
|
 |
|
  F. H. Edwards The STS Database at 20 Years: A Tribute to Dr Richard E. Clark Ann. Thorac. Surg., January 1, 2010; 89(1): 9 - 10. [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 |
