Fifteen-Year Outcome Trends for Valve Surgery in North America

Transcription

1 Fifteen-Year Outcome Trends for Valve Surgery in North America Richard Lee, MD, Shuang Li, MS, J. Scott Rankin, MD, Sean M. O Brien, PhD, James S. Gammie, MD, Eric D. Peterson, MD, Patrick M. McCarthy, MD, and Fred H. Edwards, MD, for The Society of Thoracic Surgeons Adult Cardiac Surgical Database Northwestern University Medical Center, Chicago, Illinois; Duke Clinical Research Institute, Durham, North Carolina; Centennial Medical Center, Vanderbilt University, Nashville, Tennessee; University of Maryland Medical Center, Baltimore, Maryland; and Shands Hospital, University of Florida, Jacksonville, Florida ADULT CARDIAC Background. Although results in valvular heart surgery may be improving, too few cases are available in most centers to quantify changes, especially for uncommon procedural categories. This study examined comprehensively national trends in valve surgery outcomes over the past 15 years. Methods. From 1993 through 2007, 623,039 valve procedures were grouped into single aortic (A), mitral (M), and tricuspid (T) operations, along with AM, MT, AT, and AMT multiple valves coronary artery bypass graft surgery. Pulmonary valve surgery was excluded. Trends in baseline characteristics were documented, and logistic regression adjusted for differences in patient profiles. Outcomes were expressed as unadjusted operative mortality, adjusted odds ratios for mortality, and a composite of mortality and major complications. Results. Single valves comprised 89% of valve surgery and multiple valves, 11%. Preoperative patient risk profiles worsened over time. Mortality rates were higher for multiple valves, but all mortality rates fell significantly over the 15 years (p <0.001). The composite of mortality and major morbidity did not improve, however, largely because of increasing pulmonary/infectious complications. Overall, cardiac etiology accounted for 54% of deaths, and pulmonary/infectious etiologies for 16%. Cardiac etiology of death fell by 16% over time, but pulmonary death and complications increased by 78% and 39%, respectively. Conclusions. Preoperative patient profiles for cardiac valve procedures have worsened over time. Risk-adjusted mortalities have fallen for all valve surgery, but remain higher for multiple valves. The finding of increasing pulmonary deaths and complications suggests that prevention and improved management of pulmonary and infectious complications could be an important focus for quality improvement. (Ann Thorac Surg 2011;91:677 84) 2011 by The Society of Thoracic Surgeons Outcomes in cardiac surgery seem to be improving [1 11]. Falling mortality rates have been documented for patients undergoing isolated aortic valve replacement [1, 12], primarily those with adverse risk profiles and advanced age [13, 14]. Declining mortality also has occurred for mitral valve surgery [15), due in part to increased application of valve repair [16, 17]. While mortality trends appear encouraging [1, 18, 19], comprehensive data are not available, especially for less common valve subgroups, and little is known of directional changes in postoperative morbidity. The purpose of this analysis was to evaluate baseline and outcome characteristics for all valve surgery patients in The Society of Thoracic Surgeons (STS) database, with the goal of defining comprehensively time trends in risk-adjusted mortality and morbidity. Accepted for publication Nov 8, Presented at the Forty-sixth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 25 27, Address correspondence to Dr Rankin, Vanderbilt University, 320 Lynnwood Blvd, Nashville, TN 37205; jsrankinmd@cs.com. Material and Methods This study was performed with a waiver of informed consent by the Duke University Institutional Review Board for the purpose of surgical quality improvement by analysis of deidentified patient cohorts in the STS data set. The STS currently records more than 90% of adult cardiac surgery in North America with well-developed variable sets and certified software systems [20, 21]. Detailed definitions for preoperative risk factors and postoperative complications have can be viewed online (available at: Data from individual centers are harvested quarterly, a series of data quality checks are performed before data aggregation into the national sample, and annual audits are obtained for randomly selected centers. Since 1993, the variables and definitions used have been fairly consistent and have changed only in minor ways. The accuracy and comparability of STS results have been confirmed by comparison with other mandatory and audited cardiac databases [22]. Drs McCarthy and Lee disclose that they have financial relationships with Edwards Lifesciences by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 ADULT CARDIAC 678 LEE ET AL Ann Thorac Surg VALVE SURGERY TRENDS 2011;91: Data Analysis The population for this study included 623,039 patients undergoing cardiac valve surgery in North America from 1993 to 2007 with or without concomitant coronary artery bypass graft surgery. The 15-year sample of patients was grouped by the seven types of valve operations: aortic (A), n 338,143; mitral (M), n 211,167; tricuspid (T), n 5,803; AM, n 39,260; AT, n 2,236; MT, n 21,056; and AMT, n 5,374. Patients undergoing pulmonary valve operations were excluded, as were those having Table 1. Trends in Preoperative Characteristics of Valve Surgery Patients Period Overall 1: : : Trends Variable (n 623,039) (n 136,071) (n 194,425) (n 292,543) p Values Age Mean, years Median, years years 15.0% 11.0% 14.8% 16.9% years 14.4% 13.6% 13.9% 15.1% Body mass index Mean, kg/m % 6.4% 9.3% 12.1% Diabetes 23.0% 18.3% 21.7% 26.1% Hypertension 62.8% 49.5% 60.4% 70.5% Dyslipid 45.4% 28.6% 38.7% 57.6% RF without D 5.4% 4.9% 5.4% 5.6% RF with D 2.2% 1.4% 2.1% 2.7% COPD 16.7% 7.0% 14.8% 22.4% Immune therapy 2.8% 2.0% 2.9% 3.2% CVD 13.2% 8.9% 12.9% 15.3% PVD 11.5% 9.9% 12.1% 11.9% Previous CABG 8.9% 7.8% 8.8% 9.4% Previous valve 7.7% 10.3% 7.5% 6.6% CHF % 59.6% 61.2% 56.3% Ejection fraction Median Mean % 10.0% 11.6% 12.5% Valve lesion AS 48.8% 47.0% 48.3% 50.0% AR 20.8% 16.8% 22.6% 21.5% MS 10.3% 13.4% 11.9% 7.8% MR 37.8% 27.9% 36.4% 43.3% TS 0.6% 0.6% 0.7% 0.4% TR 8.1% 3.6% 6.3% 11.3% Status Salvage 0.4% 0.9% 0.4% 0.2% Emergent 2.2% 3.0% 2.2% 1.9% Urgent 25.1% 18.2% 23.1% 29.6% Elective 71.8% 76.5% 74.0% 68.2% Endoc active 2.5% 2.1% 2.2% 2.9% Endoc treated 2.7% 2.5% 2.6% 2.8% , 2-Vessel CAD 27.6% 26.4% 26.6% 28.8% Vessel CAD 22.7% 19.5% 22.4% 24.3% LM CAD 7.6% 6.0% 7.1% 8.6% Conc CABG 45.9% 45.4% 47.7% 45.0% Baseline characteristics of the overall valve surgery patients and for the three 5-year time intervals. A aortic; CABG coronary artery bypass graft surgery; CAD coronary artery disease; CHF congestive heart failure; Conc concomitant; COPD chronic obstructive pulmonary disease; CVD cerebrovascular disease; D dialysis; Endoc endocarditis; LM left main; M mitral; PVD peripheral vascular disease; R moderate/severe regurgitation; RF renal failure; S stenosis; T tricuspid.

3 Ann Thorac Surg LEE ET AL 2011;91: VALVE SURGERY TRENDS other major concomitant procedures, such as aortic or left ventricular aneurysm repair. Patients with concomitant atrial fibrillation undergoing various ablation procedures were included. Demographic and outcome data for each of the seven groups were analyzed with individual multivariable logistic regression analyses for each operation, using standard approaches [1]. Trends in specific outcomes were documented in several ways: (1) Unadjusted operative mortality (UOM) rates were displayed on a yearly basis to allow general visualization of trends in relative outcomes for the seven valve procedures over time. Then, using each regression model, raw mortality data were normalized for differences in baseline characteristics and expressed as adjusted operative mortality. Using the middle year of the study (2000) as the reference for expected mortality, yearly observed to expected (O/E) adjusted mortality ratios were calculated for all seven procedures. (2) Because yearly event rates for less common valve procedures could be small, final statistical analysis was performed on procedural groups aggregated into three 5-year time periods: 1 (1993 to 1997), 2 (1998 to 2002), and 3 (2003 to 2007). In this analysis, operative mortalities adjusted for differences in patient baseline characteristics were expressed as adjusted odds ratios for mortality (AORM), and were documented for the seven procedures over the three time intervals. (3) Similarly, unadjusted incidences of major postoperative morbidities (neurologic defect, various pulmonary and infectious complications, and renal/ multiorgan failure) were recorded for all seven procedures over the 15 years. Similar statistical approaches as in (1), above, were used to generate yearly adjusted O/E ratios for the composite of mortality and morbidity referenced to the middle year. Risk-adjusted odds ratios for mortality and major morbidity composite (AORC) also were calculated, and evaluated over the 15 years, as in (2), above. (4) Finally, the primary cause of death, namely, the single factor that initiated complications leading to postoperative death, was assessed over time for the seven procedural categories, along with the effects of complications on mortality. Valve repair versus replacement was not included in the models for reasons of complexity associated with analyzing two operations for each of one to three valves for the seven valve procedures. 679 Results Trends in Patient Demographics Over the 15 years, demographics changed significantly (Table 1). Average age rose slowly from 66.1 to 67.2 years, and percentages of younger patients (50 to 60 years) and older patients (more than 80 years) both increased. Currently, nearly 17% of valve surgery is performed in patients more than 80 years of age. Twelve percent of valve patients have a body mass index above 35, and this incidence has doubled over the last decade. Diabetes has increased by 7%, and hypertension by more than 20%. Preoperatively, valve patients now are more likely to have renal failure (8.3%), severe lung disease (22.4%), and cerebrovascular disease (15.3%). Patients are more likely to have had prior coronary artery bypass graft surgery (9.4%), but less likely to have had prior valve surgery (6.6%). They are less likely to have congestive heart failure (56.3%) and have higher ejection fractions (median 0.55) overall. However, the percentage of patients with ejection fractions less than 0.35 is increasing and now accounts for 12.5% of the population. The type of valve disease also is changing. Degenerative disorders such as aortic stenosis are increasing, and mitral stenosis is decreasing. Aortic insufficiency, mitral insufficiency, and tricuspid insufficiency all are on the rise. The incidence of coronary artery disease is increasing, and nearly half of all valve patients now require coronary artery bypass graft surgery. Elective surgical referral continues to fall, nonelective urgent operations are increasing ADULT CARDIAC Table 2. Trends in Unadjusted Mortality (UOM) and Adjusted Odds Ratios for Mortality (AORM) (Period 1 as the Reference) for the Seven Valve Operations Period: Procedure Overall 1: : : n UOM n UOM n UOM AORM n UOM AORM p Value a All 623, % 136, % 194, % 292, % A 338, % 76, % 105, % , % M 211, % 46, % 67, % , % T 5, % 1, % 1, % , % AM 39, % 8, % 12, % , % MT 21, % 3, % 5, % , % AT 2, % % % , % AMT 5, % % 1, % , % a The p values are for the adjusted comparisons. A aortic; AM aortic-mitral; AMT aortic-mitral-tricuspid; AT aortic-tricuspid; M mitral; MT mitral-tricuspid; n number of procedures; T tricuspid.

4 ADULT CARDIAC 680 LEE ET AL Ann Thorac Surg VALVE SURGERY TRENDS 2011;91: (29.6% at present), but emergency and salvage operations are becoming less common. Trends in Mortality Average UOM for all types of valve surgery declined over time and is now 5.6% (Table 2). In particular, mortality decreased for isolated aortic and mitral valves, and for all combinations of multiple valves. The UOM after T was the highest of single valves and also fell, although not statistically significantly. The UOM for multiple valve procedures, while decreasing over the 15 years, remains more than twice as high as for single valves (Fig 1A, Table 2). After adjustment for differences in patient baseline characteristics, mortality O/E ratio and AORM declined significantly over time for all single and multiple valve procedures (Fig 1B, Table 2). The magnitude of decrease in AORM for multiple valves was similar to that of single valves. Trends in Morbidity Postoperative morbidities generally fell or remained stable over the 15 years, except for pulmonary and infectious complications, which increased significantly (Table 3). This finding was due largely to increases in prolonged ventilation, pneumonia, and multisystem organ failure. As a result, the unadjusted composite of mortality and major morbidity increased steadily over time for all valve procedures (Fig 2A). Even after adjustment for worsening risk factors, mortality O/E ratio and AORC for each procedure remained close to 1 or rose slightly over the 15 years (Fig 2B, Table 4). Patients undergoing multiple valve surgeries were more likely to experience complications. Unadjusted composite outcome worsened over time, and in the last year of the analysis (2007), the observed composite morbidity and mortality was as follows: A 20%, M 24%, T 32%, AM 36%, AT 39%, MT 35%, and AMT 46%. This phenomenon seemed to occur de- Fig 1. (A) Unadjusted mortalities and (B) observed to expected (O/E) adjusted mortality ratios for seven single-valve and multiplevalve procedures over 15 years. The middle year of the study (2000) was used as the expected reference for the O/E ratios. (Blue circles aortic [A]; pink circles mitral [M]; green circles tricuspid [T]; aqua squares A M; blue squares A T; red squares M T; black triangles A M T.)

5 Ann Thorac Surg LEE ET AL ;91: VALVE SURGERY TRENDS Table 3. Trends in Unadjusted Major Complications for Valve Surgery Over the Three 5-Year Periods Complication (%) Overall Period 1 Period 2 Period 3 (Total Group n) (623,039) (136,071) (194,425) (292,543) p Value ADULT CARDIAC Any neurologic 4.4% 4.6% 5.1% 3.9% New renal failure 7.1% 6.3% 7.1% 7.4% Multiorgan failure 2.1% 1.8% 2.1% 2.1% 0.02 Prolonged ventilation 14.5% 11.9% 13.3% 16.5% Pneumonia 4.5% 3.5% 4.5% 4.9% Sepsis 2.5% 2.3% 2.5% 2.5% Composite M/M a 22.4% 20.6% 22.0% 23.5% a Raw unadjusted values for sum of mortality and all morbidity (M/M). Morbidities are raw unadjusted data for each complication. Period to 1997, period to 2002, and period to Any neurologic complication includes stroke, encephalopathy, and transient defects. spite falling mortality rates, and largely because of increasing pulmonary/infectious complications. Trends in Causes of Death The majority of deaths were initiated by cardiac factors for all procedures and time periods (Table 5). However, a reduction in cardiac mortality occurred for all valve surgery over the 15 years, decreasing from 61% of deaths in the first period to 51% in the last. Of further interest was an increase in mortality from pulmonary and infectious etiologies, increasing from 11% in the first period to 20% in the last. Pulmonary complications Fig 2. (A) Unadjusted composite of mortality and major morbidity for the seven procedures over 15 years. (B) The observed to expected (O/E) adjusted mortality ratio for the same procedures and time with the year 2000 as the reference. (Blue circles aortic [A]; pink circles mitral [M]; green circles tricuspid [T]; aqua squares A M; blue squares A T; red squares M T; black triangles A M T.)

6 ADULT CARDIAC 682 LEE ET AL Ann Thorac Surg VALVE SURGERY TRENDS 2011;91: Table 4. Trends in Unadjusted and Adjusted Composite of Mortality and Complications for the Seven Valve Operations (Period 1 as the Reference) Period: Procedure 1: : : Overall n n n UORC AORC n UORC AORC p Value A 338,143 76, , , M 211,167 46,000 67, , T 5,803 1,000 1, , AM 39,260 8,469 12, , MT 21,056 3,328 5, , AT 2, , AMT 5, , , Trends in unadjusted odds ratio (UORC) and adjusted odds ratio (AORC) for composite mortality/morbidity for each of the seven valve operations over the three time periods. A aortic; AM aortic-mitral; AMT aortic-mitral-tricuspid; AT aortic-tricuspid; M mitral; MT mitral-tricuspid; N number of procedures; T tricuspid. consistently comprised the second most common cause of death, and were the only category that was increasing relative to the others. Occurrence of pulmonary complications increased the risk of mortality threefold to sevenfold, depending on the procedural category (Table 6). Comment The STS database has the advantage of excellent sample size, but studies can be limited by the detail of variables collected. Recently, the valve data set has been expanded, but many important aspects of patient characteristics and perioperative care are not available in the present analysis. Additionally, the data are viewed from quite a distance from each patient, and undefined confounding variables or treatment selection biases can complicate the interpretation of results. Thus, like most database studies, the findings of this paper should be interpreted within the context of observational design. However, one strength of this type of database is the ability to identify longitudinal trends over time, and for the purposes of this study, observational limitations may be less important. Demographics of North American valve patients are worsening (Table 1). Patients are increasingly more complex, sicker, and at higher baseline risk, reflecting societal trends in diabetes, hyperlipidemia, hypertension, and coronary disease [24 29]. Comorbidities are greater, and cohorts greater than 80 years of age and with severe left ventricular dysfunction are increasing. Valve reoperation after prior isolated coronary surgery is more common [30 32]. A growing population of low risk patients also was evident, possibly reflecting earlier surgery for mitral and other severe valve disease [1, 23]. Within the various valve groups, isolated tricuspid patients were younger and had more advanced cardiac and systemic disease, renal dysfunction, endocarditis, and prior cardiac surgery, seeming somewhat unique in the valve population. Aortic valve patients were older and had more coronary disease. Multiple valve patients had more comorbidities, heart failure, and reoperation but less coronary disease. The overall incidence of valve reoperation after previous valve surgery is falling, perhaps owing to improved stability of newer valve types or lower reoperation rates after valve repair [33]. Despite worsening risk profiles, UOM decreased over time for every valve category (Fig 1A). Major independent improvements in AORM also were evident over the 15 years and across all valve categories (Fig 1B, Table 2), suggesting real outcome improvement. The UOM for A procedures now averages 4.4%, and UOM for M operations has fallen to 5.8%, across all levels of risk (including emergencies, elderly patients, and so forth). Although this analysis did not separate M replacement from repair, AORM for M repair is approximately half of replacement, so transition to repair is one explanation for improved results [23]. Multiple-valve UOM ranges from 9% to 13%, lowest for MT and highest for triple valves. Although multiple-valve mortalities are falling, this area could be a specific focus for future quality improvement. Recent gains in mortality seem to have been offset by increases in morbidity, primarily pulmonary and infectious complications (Fig 2, Tables 3 and 4). Higher com- Table 5. Trends in Causes of Death After Valve Operations, Overall and by Procedure Type Cardiac Pulmonary Infection Neurologic Renal Vascular Overall (n 623,039) 53.5% 9.9% 6.6% 7.0% 3.2% 1.5% (n 136,071) 60.7% 5.3% 5.9% 6.7% 3.1% 1.1% (n 194,425) 51.3% 9.8% 6.2% 7.1% 3.3% 1.4% (n 292,543) 50.9% 12.7% 7.3% 7.0% 3.2% 1.8%

7 Ann Thorac Surg LEE ET AL 2011;91: VALVE SURGERY TRENDS Table 6. Odds Ratios for Operative Mortality for Patients With Any of Pneumonia, Septicemia, or Prolonged Ventilation (Yes Versus No) Adjusting for Other Variables Unadjusted Adjusted 683 ADULT CARDIAC Odds Ratio and 95% CI Odds Ratio and 95% CI Procedure Total n OR Lower Upper p Value OR Lower Upper p Value Aortic 338, Mitral 211, Tricuspid 5, A M 39, A T 2, M T 21, A M T 5, A aortic; CI confidence interval; M mitral; OR odds ratio; T tricuspid. plication rates may be due in part to sicker patients at baseline, but the phenomenon is evident even after adjusting for changes in baseline risk using AORC. Multiple valve outcomes were the worst, and even in 2007, more than 1 in 3 patients having multiple valves had a major morbid or fatal event. Perhaps longer cardiopulmonary bypass and aortic cross-clamp times required for multiple-valve procedures produce more patient injury, or other factors may be operative. It is not clear why multiple-valve mortality should be higher than single valves, as patient demographics and risk factors are not that different. Over the 15 years of this series, cardiac etiology has been the most important cause of postoperative death but is declining over time. This finding may be due to improvements in myocardial protection, arrhythmia management, and mechanical support, as well as better surgical procedures. The most striking finding both in the morbidity and cause of death analyses was the increase in pulmonary and infectious complications. This observation may reflect national trends of worsening antibiotic-resistant gram negative pneumonia across all of medicine [33 35]. Based on the results of this study, an effective campaign to prevent and treat pulmonary complications after valve surgery, including continued emphasis on reducing ventilator-associated pneumonia and improved management of postoperative immune dysfunction, would be a prime candidate for outcome improvement. Little emphasis has been placed on this topic in recent literature, but investigations of methods for reducing postoperative pulmonary complications are in order. In conclusion, valve surgery patients in North America exhibit worsening baseline risk profiles over time. Despite this finding, operative mortalities are steadily falling, whether assessed as raw values or risk-adjusted data. Multiple-valve surgery is associated with twice the risk of single valves and is an area for outcome improvement. Increasing use of valve repair, earlier surgical referral, and especially, reduction in pulmonary/ infectious complications are potential quality improvement candidates for future practice. References 1. Rankin JS, Hammill BG, Ferguson TB, et al. Determinants of operative mortality in valvular heart surgery. J Thorac Cardiovasc Surg 2006;131: Gillinov AM, Blackstone EH, Cosgrove DM, et al. Mitral valve repair with aortic valve replacement is superior to double valve replacement. J Thorac Cardiovasc Surg 2003;125: Panda BR, Shankar R, Kuruvilla KT, et al. Combined mitral and aortic valve replacement for rheumatic heart disease: fifteen-year follow up and long-term results. J Heart Valve Dis 2009;18: Maleszka A, Kleikamp G, Zittermann A, et al. Simultaneous aortic and mitral valve replacement in octogenarians: a viable option? Ann Thorac Surg 2008;86: David TE, Armstrong S, Maganti M, et al. Clinical outcomes of combined aortic root replacement with mitral valve surgery. J Thorac Cardiovasc Surg 2008;136: Talwar S, Mathur A, Choudhary SK, et al. Aortic valve replacement with mitral valve repair compared with combined aortic and mitral valve replacement. Ann Thorac Surg 2007;84: Sakamoto Y, Hashimoto K, Okuyama H, et al. Long-term results of triple-valve procedure. Asian Cardiovasc Thorac Ann 2006;14: Han QQ, Xu ZY, Zhang BR, et al. Primary triple valve surgery for advanced rheumatic heart disease in mainland China: a single-center experience with 871 clinical cases. Eur J Cardiothorac Surg 2007;31: Nowicki ER, Birkmeyer NJ, Weintraub RW, et al. Multivariable prediction of in-hospital mortality associated with aortic and mitral valve surgery in northern New England. Ann Thorac Surg 2004;77: Song HK, Diggs BS, Slater MS, et al. Improved quality and cost-effectiveness of coronary artery bypass grafting in the United States from 1988 to J Thorac Cardiovasc Surg 2009;137: Maganti M, Rao V, Brister S, et al. Decreasing mortality for coronary artery bypass surgery in octogenarians. Can J Cardiol 2009;25:e Brown JM, O Brien SM, Wu C, et al. Isolated aortic valve replacement in North America comprising 108,687 patients in 10 years: changes in risks, valve types, and outcomes in the Society of Thoracic Surgeons National Database. J Thorac Cardiovasc Surg 2009;137: Bhudia SK, McCarthy PM, Kumpati GS, et al. Improved outcomes after aortic valve surgery for chronic aortic regurgitation with severe left ventricular dysfunction. J Am Coll Cardiol 2007;49:

8 ADULT CARDIAC 684 LEE ET AL Ann Thorac Surg VALVE SURGERY TRENDS 2011;91: Ailawadi G, Swenson BR, Girotti ME, et al. Is mitral valve repair superior to replacement in elderly patients? Ann Thorac Surg 2008;86: Bolling SF, Dickstein ML, Levy JH, et al. Management strategies for high-risk cardiac surgery: improving outcomes in patients with heart failure. Heart Surg Forum 2000;3: Savage EB, Ferguson TB, DiSesa VJ. Use of mitral valve repair: analysis of contemporary United States experience reported to the Society of Thoracic Surgeons National Cardiac Database. Ann Thorac Surg 2003;75: Kang DH, Kim JH, Rim J, et al. Comparison of early surgery versus conventional treatment in asymptomatic severe mitral regurgitation. Circulation 2009;119: Hassan A, Quan H, Newman A, et al. Outcomes after aortic and mitral valve replacement surgery in Canada: 1994/95 to 1999/2000. Can J Cardiol 2004;20: Goodney PP, Siewers AE, Stukel TA, et al. Is surgery getting safer? National trends in operative mortality. J Am Coll Surg 2002;195: Edwards FH, Clark RE, Schwartz M. Practical considerations in the management of large multiinstitutional databases. Ann Thorac Surg 1994;58: Edwards FH, Grover FL, Shroyer AL, et al. The Society of Thoracic Surgeons National Cardiac Surgery Database: current risk assessment. Ann Thorac Surg 1997;63: Grover FL, Edwards FH. Similarity between the STS and New York State databases for valvular heart disease. Ann Thorac Surg 2000;70: Gammie JS, Sheng S, Griffith BP, et al. Trends in mitral valve surgery in the United States: results from the Society of Thoracic Surgeons Adult Cardiac Surgery Database. Ann Thorac Surg 2009;87: Catenacci VA, Hill JO, Wyatt HR. The obesity epidemic. Clin Chest Med 2009;30:415 44, vii. 25. Apovian CM. The causes, prevalence and treatment of obesity revisited in 2009: what have we learned so far? Am J Clin Nutr 2009 Nov 11 [E-pub ahead of print]. 26. Reisin E, Jack AV. Obesity and hypertension: mechanisms, cardio-renal consequences, and therapeutic approaches. Med Clin North Am 2009;93: Campbell R. Type 2 diabetes: where we are today: an overview of disease burden, current treatments, and treatment strategies. J Am Pharm Assoc 2009;49(Suppl 1): Baker JL, Olsen LW, Sorensen TI. Childhood body-mass index and the risk of coronary heart disease in adulthood. N Engl J Med 2007;357: Bibbins-Domingo K, Coxson P, Pletcher MJ, et al. Adolescent overweight and future adult coronary heart disease. N Engl J Med 2007;357: Gruberg L, Mercado N, Milo S, et al. Impact of body mass index on the outcome of patients with multivessel disease randomized to either coronary artery bypass grafting or stenting in the ARTS trial: the obesity paradox II? Am J Cardiol 2005;95: Lavie CJ, Milani RV, Ventura HO. Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss. J Am Coll Cardiol 2009;53: Gammie JS, O Brien SM, Griffith BP, et al. Influence of hospital procedural volume on care process and mortality for patients undergoing elective surgery for mitral regurgitation. Circulation 2007;115: Rankin JS, Glower DD, Teichmann TL, et al. Immunotherapy for refractory pulmonary infection after adult cardiac surgery: Immune dysregulation syndrome. J Heart Valve Dis 2005;14: Rankin JS, Burrichter CA, Walton-Shirley MK, et al. Trends in mitral valve surgery: a single practice experience. J Heart Valve Dis 2009;18: Filsoufi F, Rahmanian PB, Castillo JG, et al. Logistic risk model predicting postoperative respiratory failure in patients undergoing valve surgery. Eur J Cardiothorac Surg 2008;34: DISCUSSION DR DAVID A. FULLERTON (Aurora, CO): The frequency of pulmonary complications is interesting, particularly those leading to death. Why do you think this such a prevalent finding? DR LEE: Well, I wish I could give you hard-cut data, but all we can offer are inferences. The population undergoing valve surgery is aging and has worsening chronic obstructive pulmonary disease, but I don t think it is enough to account for this dramatic increase in pulmonary complications. Obviously, I think we have gotten better at getting people s hearts through surgery, but believe we have opened up a Pandora s box in terms of infectious etiologies. DR FULLERTON: What were the indications for surgery? For instance, were many of these cases of endocarditis, that sort of thing? DR LEE: We have. We haven t finished that analysis yet. There were a measurable number of endocarditis patients, especially in the tricuspid group. About 12% of the patients undergoing isolated tricuspid valve surgery had endocarditis, but less than 3% of the other single valves. In the multiple-valve group, only about 5% had endocarditis, but it was higher in the aortictricuspid group. I think the endocarditis accounts for some of the variability in the tricuspid population, but it still was present in the minority of patients; it does not account for the entire disparity or the differences in baseline characteristics and outcome. DR WALTER MERRILL (Jackson, MS): In terms of the patients having these difficult pulmonary problems postoperatively, does that mean that either we need to get better with their perioperative management or better in selecting the patients that we operate on? DR LEE: Yes. I think there are opportunities in each of these areas. DR NICHOLAS AUGELLI (Davenport, IA): I have a comment. One thing we noticed in our experience is the amount of blood transfusions, and there are some recent reports that indicate that these can be damaging to our older population. Do you think there is a connection between the infection and the pulmonary status and maybe the SIRS component? DR LEE: I actually do, but I can only answer with anecdotal information just like yourself. I think we need to try to tease it out, and I am not sure we can tease out the specific effect. However, I really think that transfusion contributes to pulmonary complications, especially the elderly population. I don t know about your practice, but I know in our practice we are more likely to transfuse the elderly patients because we prefer to keep their hemoglobin up, and certainly there have been multiple studies to show that transfusion encourages a systemic response, and the older the blood is, the more likely it is to cause a problem. I think that is right, but I really need to give you data before I can answer that fully.