Does Body Mass Index Affect Outcomes for Aortic Valve Replacement Surgery for Aortic Stenosis?

Transcription

1 ADULT CARDIAC Does Body Mass Index Affect Outcomes for Aortic Valve Replacement Surgery for Aortic Stenosis? Robert L. Smith II, MD, Morley A. Herbert, PhD, Todd M. Dewey, MD, William T. Brinkman, MD, Syma L. Prince, RN, William H. Ryan, MD, and Michael J. Mack, MD Cardiopulmonary Research Science and Technology Institute, and Medical City Dallas Hospital, Dallas, Texas Background. Obesity is a worldwide healthcare concern, and its association with several chronic diseases is well documented. However, the effect obesity may have on the acute care delivery is not well understood, and in cardiac surgery, reports are conflicting. The purpose of this study is to evaluate the effect of obesity in an isolated aortic valve replacement population. The hypothesis is that increasing body mass index (BMI) will portend worse long-term outcomes and greater shortterm resource utilization secondary to perioperative complications but will not affect perioperative mortality. Methods. Data were collected on 1,066 patients undergoing isolated AVR between January 2000 and December All definitions follow The Society of Thoracic Surgeons guidelines. Body mass indexes were calculated and used both as a continuous independent variable and to categorize patients into three BMI groups. Long-term mortality follow-up was by Social Security Death Index search. Standard bivariate and multivariate comparisons were performed with hierarchical models used for odds ratios. Results. When controlling for standard covariates that negatively impact outcome (sex, age, renal failure needing dialysis, diabetes mellitus, and current smoker), BMI was not predictive for either operative mortality or a composite morbidity-mortality outcome. When divided into three equal-sized groups, there was again no statistical difference among groups for mortality or for the composite variable. Separate analyses for females and males yielded the same lack of correlation. Long-term follow-up out to 12 years shows that the low BMI group has statistically worse survival than the moderate or high BMI groups. Conclusions. Increasing BMI has no independent association with worsened outcomes in the short or long term, and overweight patients have a survival benefit after surgery. Patients who are at the lower end of the BMI scale, however, are at increased risk for poor longterm survival. (Ann Thorac Surg 2012;93:742 7) 2012 by The Society of Thoracic Surgeons Obesity is a significant public health concern in the United States and worldwide. The association of obesity with several chronic disease states is well documented, as is its association with an increase in all-cause mortality [1 3]. However, in the realm of acute care delivery, the impact of obesity on outcomes is not as well understood. In the critical care literature, there is an often described U-shaped association with respect to mortality and body mass index (BMI) indicating excess mortality for those at the extremes of weight (severely underweight and morbidly obese), but a possible protective effect in the overweight and moderately obese patients [4, 5]. The enhanced survival of the obese patients is referred to as the obesity paradox, and this same phenomenon was observed in studies evaluating the effect of BMI on coronary artery revascularization outcomes [6, 7]. However, this is not a routine observation, and other studies Accepted for publication Nov 10, Presented at the Fifty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Orlando, FL, Nov 3 6, Address correspondence to Dr Smith, Cardiopulmonary Research Science and Technology Institute, Medical City Dallas Hospital, 7777 Forest Ln, Ste C-742, Dallas, TX 75230; robertl.smith@baylorhealth.edu. have demonstrated increased or no association with poor outcomes in patients with elevated BMI [8]. When evaluating the effect of BMI on aortic valve replacement (AVR) surgery, the data are similar. Rather consistently, the lower BMI patients fare poorly, but patients with an elevated BMI, despite the association with disease states diabetes mellitus, hypertension, coronary artery disease, and so forth that are associated with increased risk from cardiac surgery, demonstrate an inconsistent association with mortality [9]. The objective of this study is to evaluate the effect of BMI in a population undergoing isolated AVR for aortic stenosis. The hypothesis is that the extremes of BMI patients will have worse outcomes and greater shortterm resource utilization owing to perioperative complications. Material and Methods Study Population The North Texas Institutional Review Board at Medical City approved this study with a waiver of consent. Data were abstracted from our Society of Thoracic Surgeons 2012 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg SMITH ET AL 2012;93:742 7 EFFECT OF BMI ON AVR SURGERY OUTCOME (STS) certified database on all patients undergoing isolated AVR for aortic stenosis between January 2000 and December Patients with concomitant procedures including coronary artery bypass graft surgery (CABG) were excluded to produce a more homogeneous population. Patients must have had aortic stenosis as their primary diagnosis. Aortic insufficiency was allowed as a concomitant diagnosis, but patients with aortic insufficiency were excluded if it was the sole diagnosis. Prosthesis selection and implantation technique were according to surgeon preference. Patients undergoing a transcatheter valve replacement were excluded. There were 1,066 patients available for analysis. The surgical approach utilized was based on surgeon preference. These included both partial and full sternotomies. Study Definitions All outcomes measurements were defined according to the STS national database. Body mass index, an anthropomorphic measurement calculated by dividing the patient s weight (in kilograms) by the patients height squared (in meters), was calculated based on the patient s admission height and weight. Although BMI can be categorized according to World Health Organization (WHO) [10] international classification into six categories, we used BMI as a continuous variable. Statistical Analysis To calculate the effect of the patient obesity (BMI) on outcomes, a hierarchical multivariate logistic regression analysis was employed to measure the effect of BMI while controlling for covariates. Covariates (age, sex, renal failure requiring dialysis and diabetes) were chosen based on association with mortality and biologic plausibility of the relationship. Long-term mortality follow-up was by Social Security Death Index search for all patients. Kaplan-Meier curves were constructed using time from AVR to mortality, or last follow-up date. The log rank test was used to compare curves. Values are expressed as means SD for continuous variables or as a percentage of the group of origin for categorical variables. The results of the logistic regression are reported as odds ratios (OR) with 95% confidence intervals (CI). All statistical analyses used SAS 9.2 (SAS Institute, Cary, NC). Results Preoperative Characteristics There were 1,066 patients identified from our STS certified database for analysis in this study. Table 1 shows the preoperative characteristics of the patients. Table 1. Preoperative Characteristics of the Patients Variable % of Patients (n 1,066) Predicted risk of mortality, % Age, years BMI, kg/m Cerebrovascular disease 13.8 (147/1,066) Male 54.6 (582/1,066) Preoperative stroke 7.3 (78/1,066) Diabetes mellitus 27.4 (292/1,066) On insulin 27.6 (80/290) Dyslipidemia 57.1 (608/1,064) Family history of CVD 28.8 (307/1,065) Hypertension 71.1 (756/1,063) Immunocompromised 2.9 (31/1,066) Infectious endocarditis 1.7 (18/1,065) Peripheral arterial disease 9.5 (101/1,065) Renal failure 4.4 (47/1,066) On dialysis 31.9 (15/47) Current/recent smoker 12.1 (129/1,064) Previous CABG surgery 15.2 (162/1,065) Previous valve surgery 3.9 (42/1,065) Arrhythmia 20.6 (220/1,066) Shock 0.5 (5/1,066) Heart failure 39.0 (415/1,065) Preoperative MI 8.5 (90/1,065) Resuscitation 0.2 (2/1,065) Ejection fraction, % Bioprosthetic valve used 80.2 (852/1,063) Operative mortality 4.3 (46/1,066) Length of stay, days Cross-clamp time, minutes Perfusion time, minutes Total postoperative ventilation time, hours 743 BMI body mass index; CABG coronary artery bypass graft surgery; CVD cardiovascular disease; MI myocardial infarction. Operative Outcomes The effect of BMI on operative mortality was studied using a hierarchical logistic regression model where surgeons were nested within hospitals. The model was controlled for covariates of age, sex, diabetes, and renal failure requiring dialysis. There were 46 operative mortalities in the study, limiting the number of simultaneous covariates. We also tested several other preoperative risk factors in a univariate model and found none that were statistically significant. Table 2 presents the odds ratios and 95% confidence intervals for their effect on operative mortality. Using complications data from the STS database, a composite variable representing major morbidities and mortality was created. (Complications used included operative mortality, sepsis, deep sternal wound infection, infected thoracotomy, coma, permanent and transient stroke, perioperative myocardial infarction, reoperation for bleeding, atrial fibrillation requiring treatment, cardiac arrest, gastrointestinal complications, heart block, multi-system organ failure, tamponade, pneumonia, prolonged ventilation, postoperative renal failure, or need for a permanent pacemaker.) With a larger event rate, we added in current smoker and heart failure as covariates. Table 3 lists the odds ratios for the morbidity-mortality dependent variable. ADULT CARDIAC

3 ADULT CARDIAC 744 SMITH ET AL Ann Thorac Surg EFFECT OF BMI ON AVR SURGERY OUTCOME 2012;93:742 7 Because males and females may have different metabolic syndrome responses, we analyzed the effect of BMI separately for males and females. For the dependent variable of operative mortality, the odds ratio for the males was 1.01 (95% CI: 0.93 to 1.09; p 0.797) and for females, it was 0.99 (95% CI: 0.92 to 1.06; p 0.768). Analysis of the data for the composite morbiditymortality outcome yielded an odds ratio of 1.00 (95% CI: 0.97 to 1.04; p 0.854) for males and 0.99 (95% CI: 0.96 to 1.03; p 0.726) for females. BMI as a Categorical Variable The World Health Organization categorizes BMI into the following categories: underweight, less than 18.5 kg/m 2 ; normal weight, 18.5 to 24.9 kg/m 2 ; overweight, 25.0 to 29.9 kg/m 2 ; class I obesity, 30.0 to 34.9 kg/m 2 ; class II obesity, 35.0 to 39.9 kg/m 2 ; and morbidly obese, 40.0 kg/m 2 or more. We divided our data into three equal-sized groups with cutpoints of kg/m 2 and kg/m 2, producing groups with 355, 356, and 355 patients in the low, moderate, and high weight classes, respectively. These BMI groups are similar to those used by the WHO, with our low group the equivalent of their normal/underweight group, our moderate group similar to their overweight group, and our high BMI group similar to their obese/ morbidly obese group. Table 4 compares the preoperative risk factors and outcomes for the three BMI groups. Using the hierarchical logistic model, the odds ratio for operative mortality was calculated for the moderate and high weight groups compared to the low group. For the moderate to low comparison, the odds ratio was 0.57 (95% CI: 0.29 to 1.13; p 0.109); and for the high to low comparison, it was 0.90 (95% CI: 0.61 to 1.32; p 0.599). For the combined morbidity-mortality composite, the moderate to low comparison odds ratio was 0.95 (95% CI: 0.74 to 1.23; p 0.709); and for high to low, the odds ratio was 1.09 (95% CI: 0.91 to 1.31; p 0.361). Long-Term Outcomes A Social Security Death Index search was performed to determine the long-term mortality. The mean follow-up was years, with the low, moderate, and high groups having follow-ups of , , and years, respectively, not a statistically significant difference (p 0.183). The Kaplan-Meier curve was constructed and is displayed in Figure 1. There is a significant difference in the survival curves of the different Table 2. Adjusted Odds Ratios for Operative Mortality Variable Change in Condition Odds Ratio (95% CI) p Value BMI (kg/m 2 ) Increase of 1 kg/m ( ) Age Increase of 5 years 1.31 ( ) Diabetes mellitus Yes versus no 2.45 ( ) Sex Female versus male 1.61 ( ) Renal failure with dialysis Yes versus no 6.90 ( ) BMI body mass index; CI confidence interval. Table 3. Adjusted Odds Ratios for Composite Morbidity- Mortality Outcome Variable groups (log rank test, p 0.001), with the low BMI group demonstrating the worst long-term survival at 10 years. Survival curves for the moderate and high BMI patients tended to track fairly close together. Separating this further, the survival curve for the low BMI group was significantly different when compared with the moderate BMI group (p 0.002) and with the high BMI group (p 0.001), whereas the moderate BMI and high BMI groups were not statistically different (p 0.382). Comment Change in Condition Odds Ratio (95% CI) p Value BMI, kg/m 2 Increase of 1 kg/m ( ) Age Increase of 5 years 1.15 ( ) Diabetes mellitus Yes versus no 1.36 ( ) Sex Female versus male 1.06 ( ) Renal failure Yes versus no 1.28 ( ) with dialysis Current smoker Yes versus no 0.95 ( ) Heart failure Yes versus no 1.56 ( ) BMI body mass index; CI confidence interval. In this study of 1,066 patients undergoing isolated AVR for aortic stenosis, there was no statistically significant effect on operative mortality attributable to BMI, nor on the composite outcome of morbidity-mortality. For longterm outcomes out to more than 12 years, the moderate and high BMI groups did demonstrate better survival whereas the low BMI patients clearly had worsened survival. Also, contrary to the hypothesis, the long-term survival for the high BMI (obese/morbidly obese) group was statistically indistinguishable from that for the moderate group (overweight). Prior studies evaluating the effect of BMI on cardiac surgical outcomes have been inconsistent and difficult to compare because of the different surgical populations (CABG versus CABG/valve versus valves) and differing classification, or groupings, of obesity by BMI. Some studies, like ours, found no increased risk for mortality in patients with increased BMI. In a recent study by Florath and colleagues [9], 1,241 patients who had an AVR with or without CABG were evaluated to determine the effect of BMI. Their results showed patients with a BMI less than 24 have an increased risk of mortality at 30 days and 6 months. However, obese patients demonstrated no increased risk, not even with a BMI greater than 40 kg/m 2. Similarly, Potapov and coworkers [11] reported the effect of BMI in 22,666 patients undergoing CABG with and without valve surgery, and found that patients with low BMI (less than 20 kg/m 2 ) had increased 30-day mortality risk. Patients with a BMI of 33 to kg/m 2 had the lowest risk of mortality. This study appears to compare most similarly with our study.

4 Ann Thorac Surg SMITH ET AL 2012;93:742 7 EFFECT OF BMI ON AVR SURGERY OUTCOME Table 4. Summary of Variables for the Three Groups Variable BMI BMI BMI p Value Number of patients Predicted risk of mortality, % Age, years BMI, kg/m Cerebrovascular disease 14.4 (51/355) 14.9 (53/356) 12.1 (43/355) Male 47.6 (169/355) 64.0 (228/356) 52.1 (185/355) Preoperative stroke 7.3 (26/355) 7.6 (27/356) 7.0 (25/355) Diabetes mellitus 15.2 (54/355) 26.4 (94/356) 40.6 (144/355) On insulin 25.0 (13/52) 21.3 (20/94) 32.6 (47/144) Dyslipidemia 47.7 (169/354) 58.4 (208/356) 65.3 (231/354) Hypertension 61.9 (218/352) 72.8 (259/356) 78.6 (279/355) Immunocompromised 3.7 (13/355) 2.2 (8/356) 2.8 (10/355) Infectious endocarditis 1.4 (5/354) 2.2 (8/356) 1.4 (5/355) Peripheral arterial disease 10.7 (38/355) 9.9 (35/355) 7.9 (28/355) Renal failure 3.7 (13/355) 4.8 (17/356) 4.8 (17/355) On dialysis 1.7 (6/355) 1.7 (6/356) 0.8 (3/355) Current/recent smoker 16.4 (58/354) 9.0 (32/355) 11.0 (39/355) Previous CABG surgery 14.9 (53/355) 17.7 (63/355) 13.0 (46/355) Previous valve surgery 5.6 (20/355) 3.4 (12/355) 2.8 (10/355) Arrhythmia 23.9 (85/355) 20.5 (73/356) 17.5 (62/355) Shock 0.6 (2/355) 0.8 (3/356) 0 (0/355) Heart failure 40.6 (144/355) 37.5 (133/355) 38.9 (138/355) Preoperative MI 7.9 (28/355) 8.5 (30/355) 9.0 (32/355) Resuscitation 0.3 (1/355) 0.3 (1/355) 0 (0/355) Ejection fraction, % Operative mortality 5.6 (20/355) 2.8 (10/356) 4.5 (16/355) Length of stay, days Cross-clamp time, minutes Perfusion time, minutes Total postoperative ventilation time, hours ADULT CARDIAC BMI body mass index; CABG coronary artery bypass graft surgery; MI myocardial infarction. Although most other studies in cardiac surgery also have appreciated the higher risk for the low BMI patient [12, 13], some have found higher risk for poor outcomes in the obese patients. In a study by Prabhakar and colleagues [8], the national STS database was utilized to evaluate increased Fig 1. Long-term survival by body mass index group. BMI on CABG outcomes. In this study, 559,004 patients having first-time CABG were evaluated. Patients with a BMI of 35 to 39.9 kg/m 2 had a moderate increase in mortality risk (OR 1.21, CI: 1.1 to 1.3) compared with the normal to mild obesity group (18.5 to 34.9 kg/m 2 ). Patients with a BMI greater than 40 kg/m 2 demonstrated an odds ratio of 1.58 (CI: 1.4 to 1.7) for increased mortality. This study did not include valve surgeries, and it is difficult to truly compare the outcomes. The differences in long-term outcomes noted by our study were also interesting in that the high BMI and moderate BMI groups compared with the low BMI group conferred improved survival. These findings are difficult to explain. In a meta-analysis by Oreopoulos and colleagues [6], the effect of obesity on short-term and longterm outcomes after CABG was assessed. Although increased BMI was protective in the short-term (30-day mortality) compared with the normal BMI group, the benefit was lost in the long term (1 to 5 years after CABG). It is difficult to make direct comparisons because, again, our study involved isolated AVR, not CABG, as was the case for the Oreopoulos study [6].

5 ADULT CARDIAC 746 SMITH ET AL Ann Thorac Surg EFFECT OF BMI ON AVR SURGERY OUTCOME 2012;93:742 7 Study Limitations There are certain limitations that pertain to this study, and deserve mention. First, this study is a retrospective analysis and is limited by all inherent biases associated with this form of analysis. Second, although BMI was used similar to studies with WHO and National Institutes of Health standards for evaluating chronic disease risks, we did not obtain waist measurements. Waist measurements may be a very necessary tool to better describe size distribution. Obviously, a person who is muscular may fit into a higher BMI category despite not having a classic obese physique. The two, BMI and waist circumference together, may better predict adverse health risks in acute care patients, such as cardiac surgical patients; BMI alone may be grossly inadequate for this. Third, this study is representative of a small geographic region. As there is known regional variability in the distribution of obesity in the United States, this can affect the ability to generalize these results. In the southern states, obesity is particularly prevalent. Managing obese patients on a more routine basis may result in treatment biases that could affect outcomes in a positive or negative direction. Fourth, the population for the study is small and may fail to detect small differences in outcomes that exist. That would be better accomplished on a national level, but the STS currently uses body surface area as an anthropomorphic measure of body size in calculating risk from aortic valve surgery, not BMI. In the STS predictive model, body surface area was associated with better model performance as assessed by the c-index (defined as area under the receiver-operating curve) [14]. Lastly, the significantly poorer long-term mortality noted in the low BMI category is unadjusted. Therefore, the conclusion that can be drawn from this is limited, especially given the differences in their baseline characteristics. Based on this study and others, there appears to be no significant role of increased BMI on adverse patient outcomes. That makes counseling patients in the preoperative setting difficult. We cannot claim that weight reduction will alter surgical outcomes. However, it appears that increased BMI, specifically relating to the more obese patient, alone should not deter, or delay, operative treatment of severe aortic stenosis. In fact, a mild to moderate degree of increased BMI, as in the moderate and high BMI groups, may be protective in the short and long term. References 1. Haslam DW, James WP. Obesity. Lancet 2005;366: Allison DB, Fontaine KR, Manson JE, et al. Annual deaths attributable to obesity in the United States. JAMA 1999;282: Hu FB, Willett WC, Li T, et al. Adiposity as compared with physical activity in predicting mortality among women. N Engl J Med 2004;351: Tremblay A, Bandi V. Impact of body mass index on outcomes following critical care. Chest 2003;123: O Brien JM, Phillips GS, Ali NA, et al. Body mass index is independently associated with hospital mortality in mechanically ventilated adults with acute lung injury. Crit Care Med 2006;34: Oreopoulos A, Padwal R, Norris CM, et al. Effect of obesity on short- and long-term mortality postcoronary revascularization: a meta-analysis. Obesity (Silver Spring) 2008;16: van Straten AH, Bramer S, Soliman Hamad MA, et al. Effect of body mass index on early and late mortality after coronary artery bypass grafting. Ann Thorac Surg 2010;89: Prabhakar G, Haan CK, Peterson ED, et al. The risks of moderate and extreme obesity for coronary artery bypass grafting outcomes: a study from The Society of Thoracic Surgeons database. Ann Thorac Surg 2002;74: Florath I, Albert AA, Rosendahl UP, et al. Body mass index: a risk factor for 30-day or six-month mortality in patients undergoing aortic valve replacement? J Heart Valve Dis 2006;15: National Institutes of Health. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report. Obes Res 1998;6(Suppl 2): Potapov EV, Loebe M, Anker S, et al. Impact of body mass index on outcome in patients after coronary artery bypass grafting with and without valve surgery. Eur Heart J 2003; 24: Reeves BC, Ascione R, Chamberlain MH, Angelini GD. Effect of body mass index on early outcomes in patients undergoing coronary artery bypass surgery. J Am Coll Cardiol 2003;42: Engelman DT, Adams DH, Byrne JG, et al. Impact of body mass index and albumin on morbidity and mortality after cardiac surgery. J Thorac Cardiovasc Surg 1999;118: O Brien SM, Shahian DM, Filardo G, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 2 isolated valve surgery. Ann Thorac Surg 2009;88(Suppl): DISCUSSION DR VINAY BADHWAR (Orlando, FL): This is very helpful information. Obviously patients who are obese and short of breath from symptomatic valvular disease find it nearly impossible to exercise and lose weight before surgery. How has the advent of minimally invasive aortic valve surgery come into your algorithm for these patients? Do you not shy away from applying a minimally invasive valve to these patients or are all they done conventionally? DR SMITH: The patient population that we presented here of the 920 patients is a complete mixed bag. I think minimally invasive valve surgery is more an issue of comfort with the technique than it is of choosing necessarily the appropriate candidate. I certainly think that in most circumstances people have shied away from that technique in the obese patient population. However, that being said, there are certainly some of us who use it as a matter of routine regardless of patient size. So I think it is impossible right now to really ferret that out as far as whether that makes a difference in the outcome. DR JOHN J. KELEMEN (Salina, KS): My question is, first, a clarification of your statistics. I have seen a lot of data showing that morbidly obese people have more health problems and a shorter life expectancy than a similar normal weight person. I

6 Ann Thorac Surg SMITH ET AL 2012;93:742 7 EFFECT OF BMI ON AVR SURGERY OUTCOME was wondering if that is actually what you were comparing your data to or were some other statistics? In my practice, I have seen more and more obese people, and especially morbidly obese people. I find myself asking the question, especially if it is not an emergent or a life-saving operation, am I really helping this patient, namely, improving their quality of life, improving their longevity? We do their heart operation, we get them through it, they go home, and they don t engage in any meaningful activities because of their body habitus and other health problems. DR SMITH: I think you bring up two really good questions. As far as weight comparison with death do these patients really have a prolonged life I think one thing that is important to notice is that in the age distribution, which was shown on the preoperative risk slide, that patients who were morbidly obese were actually relatively younger than the patients who were underweight or even normal weight. So I think these patients oftentimes present at a much younger age probably because they do become so symptomatic because of their other chronic issues. So they have potentially much more life to gain as a result of an operation. This study can t answer that question. It is not designed to. As far as providing an increased life span for those patients by providing them the surgery, though, while some patients certainly may go on to live a very sedentary lifestyle, one of the problems of a lifetable or a Kaplan-Meier curve is it can t tell us how many of those patients then went on to become very active and lose significant amounts of weight and then potentially have a better outcome, and that may provide why they stayed on some sort of normal curve with the normal population. I think that is a very difficult thing to answer with this for sure, and I think a long-term study is notable to see do those patients who have aortic valve replacement surgery who are morbidly obese have higher activity levels after surgery because now they feel better. DR THORALF SUNDT (Rochester, MN): This is a terrific dataset, and I am interested in the issue of patient-prosthesis mismatch (PPM) in these obese people. Does patient-prosthesis mismatch have the same meaning in the morbidly obese person as it does in the normal weight person? You may be able to answer that question with your dataset. Have you looked at patient-prosthesis mismatch in each of these groups? DR SMITH: We did not specifically look at that, no, sir. However, I can tell you that we did look at valve size for each group, and there was no statistically significant difference in the valve size implanted for each of the groups. They all were right around between 21 to 23, as you would anticipate. And there was certainly no difference in outcomes from the morbidly obese to normal group or underweight to normal group, but in the underweight category, potentially we are having PPM in all those other groups and not in the underweight category. But we did not look at that and that would be something certainly worth going back to check out. DR SUNDT: That implies that PPM is not as much an issue in the obese patient. It might be an interesting addition to your study analysis. It implies that PPM needs to be defined maybe by ideal body mass. That is a great study. DR SMITH: We agree with that completely. 747 DR ERIC ROSELLI (Cleveland, OH): Congratulations. I think this is a really important question to ask. As we have seen more and more now with TAVI experience, understanding patients risks is something that we are falling short on. I know your groups in Dallas do an outstanding job, so don t take this the wrong way, but I was surprised to see that the normal weight patients had a quite high mortality of 6.5%, and that was your comparison group. Using such a group as your comparison group will obviously make it difficult to find a difference. So I was wondering if there was another way, perhaps the use of propensity matched analysis or something, to better compare these groups of patients more equally and are you considering changing your statistical technique? DR SMITH: We kind of battle back and forth, actually, regarding using a propensity matched analysis for doing this, and we have considered adding that as well. I think one of the things is there is a small number of outcomes and we defined a very limited set of covariates to use for our analysis based on historically significant covariates for our population. So I think from that standpoint they have worked very well, but I think it is certainly worth looking at. The STS predicted risk of mortality in these patients was highest in the underweight category and then, secondly, the normal and morbidly obese categories was the second highest risk of predicted risk mortality. I don t have those numbers in here, but they should be coming up in the manuscript. DR FAISAL BAKAEEN (Houston, TX): Great presentation. We have much more data regarding obesity and surgical outcomes in CABG patients. There was a nice study by Shroyer and colleagues published in the Annals in 2007 that shows a bimodal risk pattern, where very thin patients and very obese patients have a high risk, and those with a BMI of 30 fare the best with regard to CABG outcomes. I think you have a great study here, but the numbers are small, especially for the very obese. There is a study from the Mayo Clinic where patients with a BMI of 50 or more who underwent cardiac surgery had a high operative mortality. I think the operative mortality was about 7%. So if a patient comes to you today with a BMI of 52, are you able to look him in the eye and tell him his risk of undergoing AVR is exactly the same as a guy who is of normal weight? DR SMITH: Good question. I think if I have a guy who comes with a BMI of 52, I am lucky to look him in the eye and glad he is in clinic, and I could not honestly say from a gut perspective that that is the case. But that is the issue with data. I am presenting what the data are, and the data suggest that he should do as well. However, our dataset is limited and it is very small in numbers, especially in the obese and underweight categories, so I would have a significant hesitation in telling him that. ADULT CARDIAC