Journal of the American Heart Association ORIGINAL RESEARCH Impact of Morbid Obesity and Obesity Phenotype on Outcomes After Transcatheter Aortic Valve Replacement Angela McInerney , MD; Gabriela Tirado-­Conte, MD; Josep Rodes-­Cabau , MD; Francisco Campelo-­Parada, MD; Jose D. Tafur Soto, MD; Marco Barbanti, MD; Erika Muñoz-­Garcia, MD; Mobeena Arif, MD; Diego Lopez, MD; Stefan Toggweiler, MD; Gabriela Veiga, MD; Anna Pylko , MD; Teresa Sevilla , MD; Miriam Compagnone, MD; Ander Regueiro, MD; Viçent Serra, MD; Manuel Carnero , MD; Juan F. Oteo , MD; Fernando Rivero, MD; Henrique Barbosa Ribeiro, MD; Leonardo Guimaraes, MD; Anthony Matta, MD; Natalia Giraldo Echavarria , MD; Roberto Valvo, MD; Federico Moccetti , MD; Antonio J. Muñoz-­Garcia, MD; Javier Lopez-­Pais, MD; Bruno Garcia del Blanco, MD; Diego Carter Campanha Borges, MD; Eric Dumont, MD; Nieves Gonzalo, MD; Enrico Criscione , MD; Maciej Dabrowski, MD; Fernando Alfonso , MD; Jose M de la Torre Hernández, MD; Asim N. Cheema, MD; Ignacio J. Amat-­Santos, MD; Francesco Saia, MD; Javier Escaned , MD; Luis Nombela-­Franco , MD BACKGROUND: There is a paucity of outcome data on patients who are morbidly obese (MO) undergoing transcatheter aortic valve replacement. We aimed to determine their periprocedural and midterm outcomes and investigate the impact of obesity phenotype. METHODS AND RESULTS: Consecutive patients who are MO (body mass index, ≥40 kg/m2, or ≥35 kg/m2 with obesity-­related comorbidities; n=910) with severe aortic stenosis who underwent transcatheter aortic valve replacement in 18 tertiary hos- pitals were compared with a nonobese cohort (body mass index, 18.5–­29.9 kg/m2, n=2264). Propensity-­score matching resulted in 770 pairs. Pre–­transcatheter aortic valve replacement computed tomography scans were centrally analyzed to assess adipose tissue distribution; epicardial, abdominal visceral and subcutaneous fat. Major vascular complications were more common (6.6% versus 4.3%; P=0.043) and device success was less frequent (84.4% versus 88.1%; P=0.038) in the MO group. Freedom from all-­cause and cardiovascular mortality were similar at 2 years (79.4 versus 80.6%, P=0.731; and 88.7 versus 87.4%, P=0.699; MO and nonobese, respectively). Multivariable analysis identified baseline glomerular filtration rate and nontransfemoral access as independent predictors of 2-­year mortality in the MO group. An adverse MO phenotype with an abdominal visceral adipose tissue:subcutaneous adipose tissue ratio ≥1 (VAT:SAT) was associated with increased 2-­year all-­cause (hazard ratio [HR], 3.06; 95% CI, 1.20–­7.77; P=0.019) and cardiovascular (hazard ratio, 4.11; 95% CI, 1.06–­15.90; P=0.041) mortality, and readmissions (HR, 1.81; 95% CI, 1.07–­3.07; P=0.027). After multivariable analysis, a (VAT:SAT) ratio ≥1 remained a strong predictor of 2-­year mortality (hazard ratio, 2.78; P=0.035). CONCLUSIONS: Transcatheter aortic valve replacement in patients who are MO has similar short-­and midterm outcomes to nonobese patients, despite higher major vascular complications and lower device success. An abdominal VAT:SAT ratio ≥1 identifies an obesity phenotype at higher risk of adverse clinical outcomes. Key Words: epicardial adipose tissue ■ morbid obesity ■ subcutaneous adipose tissue ■ transcatheter aortic valve replacement ■ visceral adipose tissue Correspondence to: Luis Nombela-­Franco, MD Instituto Cardiovascular, Hospital Universitario Clínico San Carlos, Instituto de Investigación Sanitaria San Carlos (IdISSC), C/Prof. Martin Lagos s/n., 28040 Madrid, Spain. E-­mail: luisnombela@yahoo.com Supplementary Materials for this article are available at https://www.ahajo​urnals.org/doi/suppl/​10.1161/JAHA.120.019051 For Sources of Funding and Disclosures, see page 12. © 2021 The Authors. Published on behalf of the American Heart Association, Inc., by Wiley. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. JAHA is available at: www.ahajournals.org/journal/jaha J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.0190511 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR were obese, with projections in the United States sug- CLINICAL PERSPECTIVE gesting that 1 in 4 adults will have severe obesity by 2030.1,2 Coupled with this, our aging population has What Is New? resulted in a concomitant increase in the number of • The number of patients who are morbidly obese patients who are obese presenting with severe aortic undergoing transcatheter aortic valve replace- stenosis and undergoing both surgical and transcathe- ment is increasing, but a paucity of data regard- ter aortic valve replacement (TAVR). A 25-­fold increase ing their outcomes remains. in the rate of patients who are obese undergoing TAVR • Patients who are morbidly obese undergoing in the United States between 2011 and 2014 has been transcatheter aortic valve replacement have reported,3 reflecting the rapidly expanding indications similar in-­hospital, 30-­day, and 2-­year mortality for TAVR encompassing patients from low to high sur- to their nonobese counterparts. gical risk profiles.4 However, patients who are obese, • Major vascular complications are, however, in- creased in the morbidly obese population and de- and in particular patients who are morbidly obese (MO), vice success is lower, mainly driven by increased are underrepresented in clinical trials, with a paucity of mean valve gradients and patient-­prosthesis mis- data regarding the long-­term outcomes after TAVR in match; adipose tissue distribution analysis using this specific population. the pre–­ transcatheter aortic valve replacement Patients who are MO present a unique set of computed tomography scans can identify a pop- challenges, with many presenting with metabolic ulation of patients who are morbidly obese with syndrome–­related comorbidities and respiratory and an adverse obesity phenotype (visceral adipose mobility problems, which may make their periproce- tissue:subcutaneous adipose tissue ratio ≥1) and dural recovery more complex. However, within the increased risk of mortality at 2 years. domain of cardiovascular disease, the presence of an “obesity paradox” continues to be debated, although What Are the Clinical Implications? recently, this paradigm has been challenged in patients • Patients who are morbidly obese can safely be offered transcatheter aortic valve replacement who are MO.5,6 In the TAVR field, conflicting results as a treatment for severe aortic stenosis. exist concerning the periprocedural and long-­term out- • Vigilance is required when performing vascu- comes in patients who are obese.7–­13 Interpretation of lar access in patients who are morbidly obese these data is further hampered by small patient num- who have an increased risk of major vascular bers, particularly in the MO group,14 involving mainly complications. single centers, and heterogeneity of body mass index • Adipose tissue distribution can identify an ad- (BMI) cutoff points used to define patient catego- verse obesity phenotype and provide important ries. Furthermore, the use of BMI is a relatively crude prognostic information beyond that of body marker of obesity being unable to differentiate between mass index alone, which may avoid futile proce- fat mass and muscle mass. Given these limitations, dures in patients who are morbidly obese. there is an increasing interest in the use of an obesity or metabolic phenotype to more clearly differentiate metabolically healthy from unhealthy patients who are obese rather than simply using BMI alone.15 Computed tomography (CT) scans can be used to assess a pa- Nonstandard Abbreviations and Acronyms tient’s obesity phenotype and assess the distribution EAT epicardial adipose tissue of adipose tissue components. The objectives of our IMAT intramuscular adipose tissue study, therefore, were (1) to compare periprocedural iVAT indexed visceral adipose tissue complications and midterm outcomes in a large cohort of patients who are MO, both unmatched and matched MO morbidly obese to a nonobese cohort; (2) to determine prognostic fac- MVC major vascular complication tors among patients who are MO after TAVR; and (3) PPM patient-­prosthesis mismatch to evaluate the predictive value of adipose tissue dis- SAT subcutaneous adipose tissue tribution on clinical outcomes in patients who are MO. TAVR transcatheter aortic valve replacement VAT visceral adipose tissue METHODS Requests to access the data set, analytic methods, W orldwide, the obesity epidemic continues to and study materials may be sent to the correspond- grow. In 2016, the World Health Organization ing author. This was a multicenter study collecting estimated that 650 million people worldwide individual data on consecutive patients who are MO J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.0190512 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR with symptomatic severe aortic stenosis who under- the muscle capsule. Areas (centimeters squared) of went TAVR in 18 tertiary care centers from Europe VAT, SAT, and IMAT, and skeletal muscle area (SMA) and the United States. In addition, 6 centers provided were measured and recorded for each slice. The full data on their entire nonobese TAVR cohort as a measurements from the 2 third lumbar spine slices comparator group. The decision to perform TAVR were averaged and all results were indexed to body was made at each center as per their local protocol, surface area (iVAT, indexed SAT, indexed IMAT, in- and TAVR was performed as previously described.16 dexed SMA; cm2/m2). An analysis of the effect of The study was performed in accordance with the in- sarcopenic obesity was also performed on the basis stitutional review board of participating centers, and of previously defined sex-­specific SMAs indexed to all patients provided informed consent for the pro- height squared (height2) (rather than body surface cedures. All procedural-­related aspects were at the area). Using this definition, height-­ indexed SMA operator’s discretion. ≤38.5 cm2/m2 for women and ≤52.5 cm2/m2 for men BMI was calculated as weight in kilograms/height were considered sarcopenic.23 The proportion of VAT in meters squared. The following definitions were ap- compared with SAT was calculated by the VAT:SAT plied to define the study groups such that the non- ratio and percentage VAT [(VAT/(VAT+SAT))*100] as obese group included patients with a BMI of 18.5 to previously described.24 A second analysis of epicar- 29.9 kg/m2, and the MO group included patients with dial adipose tissue (EAT), defined as all fat contained a BMI ≥40 kg/m2, or ≥35 kg/m2 with obesity-­related within the fibrous pericardium from the bifurcation of comorbidities.17,18 Patients with a BMI <18.5 kg/m2 or the pulmonary artery to the diaphragm,25 was per- 30 to 34.9 kg/m2 were excluded. Baseline, periproce- formed on contrast cardiac CT studies. A Hounsfield dural, and follow-­up clinical data were prospectively unit range of −190 to −30 was used.25 EAT was man- collected in a dedicated database in each participat- ually redefined every 3 slices, with interpolation of ing center and the coordinating center performed the intervening slices. The software calculated the EAT statistical analysis. Periprocedural events were defined volume (centimeters cubed) by summing the EAT using the Valve Academic Research Consortium-­2 cri- area in each slice and taking into account slice thick- teria, including device success and patient-­prosthesis ness and intersection gap (Figure S1 and Video S1). mismatch (PPM) with differing cutoff points for obese Results were also indexed to body surface area (cen- and nonobese groups as specified.19 To determine timeters cubed/meters squared). the presence of PPM, previously defined predicted ef- fective orifice area for each valve type and size were Statistical Analysis used.20 Clinical follow-­up was at 30 days, 12 months, Categorical variables were expressed as number and yearly thereafter. Midterm outcomes were as- and percentage while continuous variables were sessed at 24 months. expressed as mean and SD or median and inter- quartile range (25th–­ 75th percentile) according to CT Analysis of Adipose Tissue their distribution. Normality was assessed using the Distribution Kolmogorov-­Smirnov text. Qualitative variables were An adipose tissue distribution substudy was per- analyzed using the χ2 or the Fisher exact test and formed on baseline pre-­TAVR CT scans. CT images differences in continuous variables using a 2-­sided from patients who are MO were centrally analyzed in t test or Wilcoxon rank test for the unmatched com- a core laboratory at the coordinating center using a parison. A propensity score–­matched analysis was specific software package (sliceOmatic version 5.0; also performed between the 2 groups. A propen- TomoVision, Quebec, Canada). Two separate analy- sity score was estimated using a logistic regression ses were performed. First, an analysis of abdominal model. Morbid obesity was the dependent variable, fat components on 2 consecutive cross-­ sectional and independent variables were those baseline slices taken at the third lumbar spine as previously characteristics found to have statistically significant described.21,22 The software package was used to differences between obesity groups and other vari- segment each slice into its various tissue components ables considered to be clinically relevant. The final based on Hounsfield units in a semiautomatic fash- variables included in the propensity matching were ion. Adipose tissue was measured using a Hounsfield age, sex, diabetes mellitus, coronary artery disease, unit threshold between −190 and −30, while muscle chronic obstructive pulmonary disease, peripheral was set as Hounsfield units between −29 and +150. vascular disease, and femoral access. A propen- Visceral (VAT) and subcutaneous (SAT) adipose tis- sity score–­matched cohort was then created with sue were separated by the borders of the abdomi- a 1:1 ratio of nonobese patients and patients who nal muscles (Figure 1). Intramuscular adipose tissue were MO using a “nearest neighbor” match with- (IMAT) was considered to be any fat contained within out replacement. A caliper of <0.1× the SD of the J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.0190513 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR Figure 1. Favorable and unfavorable obesity phenotype. Abdominal adipose tissue distribution from a single computed tomography slice is shown with subcutaneous (blue) and visceral (yellow) adipose tissue, separated by the abdominal muscular layer (red). Intramuscular adipose tissue is seen in green. Favorable metabolic phenotype is seen in a female patient with a BMI of 44 kg/m2 and a visceral to subcutaneous adipose tissue ratio (VAT:SAT ratio) of 0.52. Unfavorable metabolic phenotype is seen in a male patient with a BMI of 40.5 kg/m2 and a VAT:SAT ratio of 1.82. BMI indicates body mass index; SAT, subcutaneous adipose tissue; and VAT, visceral adipose tissue. logistic score was applied. Standardized differences variable) were associated with 2-­year all-­c ause mor- were calculated for all covariates before and after tality. Cox regression analysis was then repeated matching and represented graphically in density exclusively in the patients who were MO to further plots to assess balance (Figure S2). Comparison evaluate independent predictive factors for all cause of continuous and categorical variables between 2-­year mortality in the MO group. Proportionality the matched groups were as previously stated for hazard assumption for Cox models was tested unmatched groups. Freedom from mortality and re- using Schoenfeld residuals. Testing for influential admission curves were calculated using the Kaplan-­ observations was with the DFBETA index. Adipose Meier method and compared using the log-­ rank tissue measurements were additionally analyzed to test in both the unmatched and matched cohorts. investigate their impact on all-­c ause and cardiovas- Cox regression analysis was used in the whole un- cular mortality and readmission rates. All data were matched cohort to determine if morbid obesity (as analyzed with Stata version 15.1 (StataCorp, College a dichotomous variable) or BMI (as a continuous Station, TX). J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.0190514 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR Table 1. Baseline Characteristics in Nonobese and Morbidly Obese Cohorts Unmatched Cohort Matched Cohort BMI 18.5–­29.9 BMI >35 BMI 18.5–­29.9 BMI >35 (n=2264) (n=910) P Value (n=770) (n=770) P Value Age, y 81.7 (6.4) 76.8 (7.4) <0.001 78.1 (7.4) 78.5 (6.2) 0.206 Female sex 1138 (50.3) 608 (66.8) <0.001 473 (61.4) 487 (63.3) 0.462 Body mass index, 25.7 (2.7) 39.5 (5.2) … 25.9 (2.7) 38.9 (4.2) … kg/m2 Diabetes mellitus 659 (29.1) 496 (54.5) <0.001 388 (50.4) 377 (49) 0.575 Insulin use 100 (24.6) 187 (40.9) <0.001 59 (25.3) 133 (38.1) 0.001 Hypertension 1845 (81.5) 850 (93.4) <0.001 635 (82.5) 727 (94.4) <0.001 Hyperlipidemia 1117 (59.9) 652 (74.7) <0.001 443 (64.2) 533 (72.6) 0.001 Baseline creatinine, 1.24 (0.79) 1.26 (0.86) 0.545 1.22 (0.87) 1.26 (0.84) 0.398 mg/dL eGFR <30 mL/ 177 (7.8) 79 (8.7) 0.419 58 (7.5) 64 (8.3) 0.571 min/1.73 m2 Coronary artery 1017 (44.9) 391 (43) 0.312 349 (45.3) 328 (42.6) 0.281 disease Previous myocardial 295 (13.1) 111 (12.3) 0.551 132 (17.2) 92 (12.0) 0.004 infarction Previous PCI 563 (24.9) 228 (25.1) 0.904 186 (24.2) 189 (24.6) 0.859 Previous CABG 173 (7.7) 111 (12.2) <0.001 71 (9.2) 94 (12.2) 0.060 Previous valve surgery 116 (5.1) 57 (6.3) 0.199 45 (5.9) 45 (5.9) 0.995 Valve-­in-­valve TAVR 103 (4.6) 50 (5.9) 0.160 41 (5.4) 41 (5.7) 0.800 Atrial fibrillation 743 (32.9) 317 (34.9) 0.269 257 (33.4) 268 (34.9) 0.542 Previous pacemaker 222 (9.8) 82 (9.1) 0.556 67 (8.7) 77 (10.0) 0.381 COPD 456 (20.2) 264 (29.0) <0.001 210 (27.3) 194 (25.2) 0.354 Previous 291 (12.9) 100 (11) 0.146 102 (13.3) 91 (11.8) 0.392 cerebrovascular accident Peripheral vascular 362 (16.0) 111 (12.2) 0.006 108 (14.0) 101 (13.1) 0.602 disease NYHA functional class 1198 (53.1) 663 (73.0) <0.001 485 (63.3) 547 (71.2) 0.001 III and IV Baseline hemoglobin, 12.0 (1.7) 12.0 (1.6) 0.615 11.8 (1.6) 12.0 (1.6) 0.050 g/dL NT-­proBNP 2154 [840–­5163] 1133 [380–­2800] <0.001 1758.5 [577.5– ­4197] 1183 [428 0.002 –­2902] Logistic EuroSCORE 13.64 [9.12–­22.34] 11.13 <0.001 12.91 [8.225–­21] 11.59 0.037 [7.01–­18.66] [7.79–­19.39] EuroSCORE II 4.1 [2.54– ­6.86] 3.58 [2.16–­5.77] 0.001 4.135 [2.5–­7.07] 3.6 [2.20– ­5.95] 0.011 STS 4.7 [3.2–­7.034] 4 [2.725–­6.104] <0.001 4.6 [2.96– ­6.951] 4 [2.765–­6.258] <0.001 Echocardiographic data LVEF 56.3 (14) 57.1 (11.2) 0.138 56.6 (14.5) 57.1 (11.1) 0.459 Mean aortic gradient, 45 [37–­56] 46 [39–­56] 0.180 45 [36–­55] 46 [39–­55] 0.273 mm Hg Aortic valve area, cm2 0.68 (0.23) 0.73 (0.20) <0.001 0.67 (0.24) 0.73 (0.21) <0.001 Moderate or severe 397 (18.9%) 135 (16%) 0.210 152 (20.0%) 112 (15.7%) 0.030 mitral regurgitation Moderate or severe 274 (12.3%) 96 (11.3%) 0.454 100 (13.2%) 83 (11.6%) 0.351 aortic regurgitation Moderate or severe 638 (36.0%) 384 (48.2%) <0.001 270 (43.1%) 316 (47.5%) 0.114 PHT Values are expressed as mean (SD), median [IQR] or n (%). BMI indicates body mass index; CABG, coronary artery bypass grafting; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; LVEF, left ventricular ejection fraction; NT-­proBNP, N-­terminal pro-­B-­t ype natriuretic peptide; NYHA, New York Heart Association; PCI, percutaneous coronary intervention; PHT, pulmonary hypertension; STS, Society of Thoracic Surgeons; and TAVR, transcatheter aortic valve replacement. J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.0190515 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR RESULTS between groups (Figures 3C and 3D). Figure S3 depicts outcomes in the unmatched cohort. Patient Population Thirty-­four (3.6%) patients with a BMI of 35 to 39.9 kg/ Predictors of Outcome in Patients Who m2 not meeting the MO definition17,18 were not included in the analysis. Finally, a total of 3174 patients undergo- Are MO ing TAVR were included; 2264 in the nonobese group Table S3 lists predictors of mortality in the whole and 910 in the MO group. Baseline characteristics of cohort, while independent predictors of 2-­year mor- the unmatched population are summarized in Table 1. tality in the MO population (n=910) are shown in Patients who were MO were younger (76.8 versus Table 4. Decreased baseline estimated glomerular 81.7 years), more likely women (66.8% versus 50.3%), filtration rate (hazard ratio [HR], 1.16; 95% CI, 1.07–­ with lower surgical risk scores (median Euroscore 1.26; P<0.001) and nontransfemoral vascular ac- II 3.58 versus 4.10; P=0.001). Procedural aspects of cess (HR, 1.82; 95% CI, 1.19–­2.79; P=0.006) were the unmatched cohorts are summarized in Table S1. both independent predictors of all-­cause mortality at Vascular access was predominantly transfemoral 2 years. Neither BMI, body surface area, nor severe (≈87%) in both groups with fewer patients who are MO PPM predicted 2-­year mortality in the whole or MO having general anesthesia. cohorts. After propensity score application, a matched cohort with 770 patients per group was obtained. Adipose Tissue Distribution Subanalysis Although close matching was achieved (Figure S2A Of 394 CT scans collected that included third lum- and B), some baseline cardiovascular risk factors, such bar spine images, 275 (69.8%) were suitable for ab- as hypertension, hyperlipidemia, and insulin-­requiring dominal VAT, IMAT, and SMA analysis. Excessively diabetes mellitus, continued to differ between groups narrowed field of view (n=85) and asymmetry (n=34) (Table 1). Most procedural aspects in the matched were the main reasons for excluding CT scans. A fur- group were similar, although general anesthesia was ther 56 patients were excluded from SAT analysis due still less common in patients who were MO (Table 2). to an excessively narrowed field of view resulting in Patients who were MO were also more likely to have 219 SAT analyses being performed. The mean area larger-­sized valves implanted (34.1% versus 28.6% for for each abdominal adipose tissue compartment in- valve sizes 29–­34 mm; P=0.020), with less postdilation dexed to body surface area were iVAT 146.0±53 cm2/ compared with the nonobese group. m2, indexed SAT 172.4±56.3 cm2/m2, and indexed IMAT 16.6±8.7 cm2/m2. The average indexed SMA In-­Hospital and Midterm Outcomes was 60.2±13.5 cm2/m2. The proportion of IMAT inside Table S2 and Table 3 summarize periprocedural com- the skeletal muscles was 22±10%. A total of 376 con- plications in the unmatched and matched populations, trast cardiac CT scans were analyzed, and epicardial respectively. Major vascular complications (MVCs) oc- fat volume quantification was feasible in 266 (70.74%). curred more commonly in the MO group (unmatched, Reasons to exclude were excessively narrowed field 6.6% versus 4.6%, P=0.019; and matched 6.6% (n=53) or difficulty in identifying the pericardial out- versus 4.3%, P=0.043). No differences in bleeding, line (n=59). The average indexed EAT volume was hospital-­acquired pneumonia, and acute kidney in- 47.0±23.1 cm3/m2. Sarcopenic obesity was found in jury, were noted. Device success was lower in the MO 7.84% of the population using predefined sex-­specific group (unmatched, 83.4% versus 89.7%, P=0.001; and cutoffs. Absolute and indexed areas and volumes of matched 84.4% versus 88.1%, P=0.038). Reasons for adipose and muscle tissue analysis are summarized not achieving device success were different between in Table S4. groups (Figure 2). Patients who were MO had higher Most periprocedural events were not associated rates of severe PPM and mean aortic valve gradients, with adipose tissue distribution parameters assessed, which remained higher at 1-­year follow-­up. No differ- except for periprocedural cerebrovascular accident, the ences were observed for short-­term mortality (Table 3 risk of which increased with increasing iVAT (odds ratio, and Table S2). 2.09 for each 40 cm2/m2 incremental increase; 95% CI, After a median follow-­up of 14.1 months (interquartile 1.05–­4.15; P=0.036). The optimal cutoff for predicting range, 6.5–­ 36.0), survival analysis in the matched periprocedural cerebrovascular events was an iVAT of cohort demonstrated similar rates of freedom from 221.3cm2/m2, giving an area under the curve of 0.802 all-­ cause and cardiovascular mortality for MO and (95% CI, 0.4949–­1.000) and a sensitivity of 66.67% and nonobese groups (79.4 versus 80.6%, P=0.731; and specificity of 91.91%. Indexed IMAT and indexed SMA 88.7 versus 87.4%, P=0.699, respectively) (Figure 3A were not associated with periprocedural or midterm and 3B). All-­ cause and cardiovascular readmission outcomes, and no association was found between CT-­ rates at 24 months were high overall, with no difference defined sarcopenic obesity and 2-­ year mortality. An J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.0190516 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR Table 2. Procedural Aspects of Propensity Score–­Matched Nonobese and Morbidly Obese Cohorts BMI 18.5–­29.9 BMI >35 (n=770) (n=770) P Value Procedural urgency Urgent/emergent procedure 63 (9.1) 50 (7.3) 0.222 Access site Transfemoral 684 (88.8) 670 (87.0) 0.274 Nontransfemoral 86 (11.17) 100 (12.99) 0.274 Method of transfemoral access*,† Percutaneous with closure device 550 (86.5) 584 (90.7) 0.018 Surgical cutdown 86 (13.5) 60 (9.3) 0.018 Prosthesis type BEV 332 (43.1) 353 (45.8) 0.282 SEV 438 (56.9) 410 (53.3) 0.151 Edwards Sapiens & XT & S3 332 (43.1) 353 (45.8) 0.282 Medtronic Corevalve, Evolut R, Evolut Pro 363 (47.1) 347 (45.1) 0.413 Other (Portico, accurate neo, other) 75 (9.7) 70 (9.1) Prosthesis size 20–­23 mm 199 (26) 170 (22.2) 0.083 25–­27 mm 346 (45.4) 328 (43.40) 0.442 29–­34 mm 218 (28.6) 258 (34.1) 0.020 General anesthesia 354 (46) 288 (37.4) 0.001 Prior balloon valvuloplasty 454 (63) 418 (60) 0.246 Balloon post-­dilatation 151 (20.4) 93 (12.3) <0.001 Values are expressed as n (%). BEV indicates balloon expandable valve; BMI, body mass index; and SEV, self-­expanding valve. *One center that practices only “cutdown” technique for femoral access was excluded. † Femoral access only. increased risk of all-­cause death at 2 years was found for each 10cm3/m2 increment in indexed epicardial adi- DISCUSSION pose tissue (HR, 1.16; 95% CI, 1.03–­1.30; P=0.011) and The main findings of the study include the following: each 40 cm2/m2 incremental increase in iVAT (HR, 1.40; (1) TAVR in patients who are MO is a safe procedure 95% CI, 1.05–­1.86; P=0.021). Furthermore, the propor- with similar periprocedural complications to a co- tion of visceral to subcutaneous fat was associated with hort of nonobese patients, except for a higher rate of 2-­year mortality: a ratio of VAT:SAT ≥1 demonstrated MVCs; (2) device success was lower in the MO co- the best cutoff point for identifying patients who were hort, mainly driven by a higher rate of elevated mean MO with an unfavorable obesity phenotype for 2-­year aortic gradient and severe PPM, but severe PPM was mortality with and an area under the receiver operat- not a predictor of 2-­year mortality in the MO group; ing characteristic curve of 0.696, giving a sensitivity of (3) similar 2-­year outcomes were observed in the un- 63.7% and specificity of 71.1%. This cut point gave an matched and matched MO and nonobese groups; HR for all-­cause mortality of 3.06 (95% CI, 1.20–­7.77; and (4) adipose tissue distribution analysis obtained P=0.019), cardiovascular mortality of 4.11 (95% CI, from the baseline CT scan identified an obesity phe- 1.06–­15.90; P=0.041), and readmission of 1.81 (95% notype (VAT:SAT ratio ≥1) at greater risk of all-­cause CI, 1.07–­3.07; P=0.027) at 2 years. Figure 1 depicts the and cardiovascular mortality and readmission at adipose tissue distribution analysis of a favorable and 2 years. Our conclusions for this study are supported unfavorable obesity phenotype. Kaplan-­Meier curves by a large study population, a robust methodology for 2-­year all-­cause and cardiovascular mortality of pa- used to measure adipose tissue distribution, midterm tients with a VAT:SAT ratio <1 and ≥1 are depicted in follow-­up, and use of propensity score matching as Figure 4. After adjustment for other factors associated part of the statistical analysis. with mortality, a VAT:SAT ratio ≥1 remained a strong Short-­and midterm survival was not different be- predictor of 2-­year mortality with an HR of 2.78 (95% tween groups in either the unmatched or matched CI, 1.07–­7.19; P=0.035). Differences in baseline charac- population in our study. These findings suggest that teristics in patients with a VAT:SAT ratio <1 and ≥1 are patients who are MO are equally likely to benefit from shown in Table S5. a TAVR procedure than their nonobese counterparts. J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.0190517 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR Table 3. Clinical End Points and Echocardiographic Data However, it does not support the notion of an “obesity After the Procedure in the Propensity Score–­Matched paradox” and is more in keeping with prior studies on Analysis for Nonobese and Morbidly Obese Cohort populations such as ours7 using BMI ≥35 kg/m2 and Clinical End BMI 18.5–­29.9 BMI >35 with other studies investigating the impact of BMI on Points (n=770) (n=770) P Value outcomes, which showed no difference in mortality In-­hospital mortality 26 (3.4) 28 (3.6) 0.782 rates between obese and normal-­weight patients.26,27 In-­hospital or 30-­d 28 (3.6) 35 (4.6) 0.368 Variations in reported results in studies to date exam- mortality ining the impact of BMI on outcomes in TAVR may be Vascular complications related to the use of a single BMI cutoff for obesity Major 33 (4.3) 51 (6.6) 0.043 (generally >30 kg/m2), which fails to take account of Minor 80 (14.1) 61 (9.3) 0.009 varying obesity phenotypes and may lack sensitivity Vascular complications femoral access only to discern differences in outcomes between obesity Major 32 (4.7) 48 (7.2) 0.052 subgroups.15 Our study concentrated on the MO sub- Minor 79 (16.1) 58 (10) 0.003 group, who unlike the moderately obese cohort (BMI, 30–­35 kg/m2) often have adverse risk factor profiles, Major vascular complications femoral access only by closure type and, despite being commonly encountered in real-­ Percutaneous 22 (4.0) 39 (6.7) 0.046 closure device world practice, are often underrepresented in clinical Surgical cutdown 8 (9.3) 9 (15.0) 0.291 trials. However, with comparable outcomes to the non- technique obese group, our findings suggest that TAVR is a viable Bleeding option in this population and should be offered to pa- Life-­threatening 22 (2.9) 19 (2.5) 0.659 tients who are MO with severe aortic stenosis. bleeding Some differences in valve hemodynamics were noted Major bleeding 37 (4.8) 44 (5.7) 0.421 between groups, which affected device success rate. Life-­threatening 59 (7.7) 63 (8.2) 0.706 Increased rate of PPM in both surgical AVR and TAVR and major has been previously documented in obese patients.28–­30 Minor bleeding 66 (8.6) 55 (7.3) 0.353 Likewise, mean aortic gradient and severe PPM was AKI more common in the MO group, despite using larger Stage I 85 (13.7) 101 (14.8) 0.548 valve sizes. These findings suggest that careful con- Stage II and III 24 (3.9) 25 (3.7) 0.859 sideration must be given to choose a prosthesis with a Any stage 109 (17.5) 126 (18.5) 0.646 favorable hemodynamic profile to optimize results in the MO population. Surprisingly, we found a decreased rate Coronary occlusion 5 (0.7) 4 (0.5) 0.738* of moderate to severe aortic regurgitation after TAVR Periprocedural 11 (1.4) 12 (1.6) 0.831 Stroke in the MO group (in both the matched and unmatched Hospital-­acquired 11 (1.57) 10 (1.33) 0.690 populations), in accordance with Sharma et al.7 Whether pneumonia this could be related to poorer visualization of regurgi- New permanent 91 (13) 108 (15.6) 0.158 tant aortic flow or additional factors that could influence pacemaker paravalvular leak should be further investigated in future implantation dedicated studies. Length of hospital 6 [5–­9] 5 [3–­8] <0.001 Similar to the nonobese group, 20% of the MO stay, d cohort died within 2 years, making careful patient se- Post-­TAVR echocardiogram parameters within 30-­d after TAVR lection to avoid futile interventions an important issue Moderate-­severe 38 (5.0) 18 (2.5) 0.009 for this population. Multivariable analysis highlighted a post-­TAVR AR number of predictive factors among the MO population Postprocedural 8 [6–­11] 10 [7–­14] <0.001 mean aortic valve for 2-­year mortality, which importantly included non- gradient (mm Hg) transfemoral access, suggesting transfemoral route Severe patient-­ 7 (1.1) 23 (3.5) 0.004 should be the preferred approach in this cohort, as in prosthesis others.4 Notwithstanding the preference for the trans- mismatch femoral route, MVCs were higher in the MO group, Device success 678 (88.1) 650 (84.4) 0.038 including in those who had a transfemoral TAVR per- Echocardiogram parameters at 1-­y after TAVR formed, a finding also reported by González Ferreiro 1 y mean aortic 8 [6–­11] 10 [7–­14] <0.001 et al14 and previously found by Hibbert et al31 in the valve gradient (mm Hg) context of femoral access coronary angiography in patients with extreme obesity. Performing safer percu- Values are expressed as n (%) or median [IQR]. AKI indicates acute kidney injury; AR, aortic regurgitation; BMI, body mass index; and TAVR, taneous procedures in patients who are MO requires transcatheter aortic valve replacement. vigilance when dealing with vascular access. Although *Fisher’s exact test used. more MVCs were seen in the MO group who had a J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.0190518 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR Figure 2. Chart depicting the causes of device failure among the matched nonobese patients vs patients who are morbidly obese. purely percutaneous approach with a closure device, which is consistent with the findings of Mancio et this method had fewer MVCs than those who under- al,35 who also found high levels of iVAT to be associ- went surgical cutdown for femoral vascular access. A ated with all-­cause mortality in obese (n=44), but not recent study by Kotronias et al32 demonstrated a signif- normal weight patients. The large number of patients icant reduction in vascular access complications with included in our study with iVAT assessment adds the use of ultrasound-­guided access. Further studies considerable weight to this finding. Importantly, we investigating the importance of preventive strategies found that the proportion of visceral to subcutane- such as contralateral vascular protection for percuta- ous fat had a strong association with readmission neous access in this challenging population are cur- and midterm all-­cause and cardiovascular mortality rently awaited (Junquera et al, under review). with a VAT:SAT ratio ≥1 resulting in a more than 3-­fold increased risk of all-­cause, and 4-­fold increased risk of cardiovascular death at 2 years. Furthermore, this Obesity Phenotype association with all-­cause mortality was maintained Adipose tissue distribution analysis can further dis- when the VAT:SAT ratio was added to our multivari- cern a patient’s obesity phenotype and has the poten- able analysis. This was despite very similar baseline tial to add another dimension to the risk stratification characteristics in patients with a VAT:SAT ratio >1 and of patients. Visceral adiposity has been found to be <1 (Table S5), suggesting an added benefit for this associated with development of cardiovascular risk ratio in predicting outcomes beyond normal cardio- factors and adverse cardiac outcomes, particularly in vascular risk factors or indeed BMI. Visceral fat, an relation to coronary artery disease,33,34 and improves ectopic fat deposit, is associated with dysregulated risk prediction for cardiovascular events beyond that fat metabolism and insulin resistance and is known of BMI alone.33 Its association with outcomes in a MO to be a highly metabolically active tissue,37 while ab- TAVR population has not been previously described, dominal subcutaneous fat has been shown to have although some small studies in non-­MO TAVR popu- an inverse relationship with atherosclerotic disease38 lations exist.35,36 Our study found an increased risk and, in the context of TAVR, an inverse relationship of all-­cause mortality at 2 years with increasing iVAT, with mortality.36 Furthermore, increasing VAT:SAT J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.0190519 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR A B 100 100 Freedom from all-cause mortality (%) Freedom from CV mortality (%) 92.4% 88.7% 87.1% 91.2% 79.4% 87.4% 75 75 87.5% 80.6% 50 50 Log rank, p=0.731 Log rank, p=0.699 25 25 Non obese Non obese Morbidly obese Morbidly obese 0 0 0 6 12 18 24 0 6 12 18 24 Time (months) Time (months) Number at risk Number at risk Non obese 763 596 516 383 330 Non obese 763 596 516 383 330 Morbidly obese 763 573 468 300 240 Morbidly obese 763 573 468 300 240 C D 100 100 Freedom from CV readmission(%) Freedom from readmission (%) 74.1% 85.4% 78.7% 75 75 86.1% 73.7% 60.8% 79.2% 60.5% 50 50 Log rank, p=0.508 Log rank, p=0.335 25 Non obese 25 Non obese Morbidly obese Morbidly obese 0 0 0 6 12 18 24 0 6 12 18 24 Time (months) Time (months) Number at risk Number at risk Non obese 763 505 416 284 227 Non obese 763 505 416 284 227 Morbidly obese 763 511 395 235 174 Morbidly obese 763 511 395 235 174 Figure 3. Kaplan-­Meier graph demonstrating 2-­year all-­cause (A) and cardiovascular (B) mortality and 2-­year all-­cause (C) and cardiovascular (D) readmission in the propensity score–­matched analysis for obese and nonobese groups. CV indicates cardiovascular. ratio has previously been shown to be an independ- strong relationship with midterm outcomes. The sim- ent predictor of cardiac events in patients referred for ilar association with percentage VAT and 2-­year mor- CT coronary angiography.39 It is unsurprising, there- tality further reinforces the notion that adipose tissue fore, that VAT:SAT ratio in our study demonstrated a distribution, rather than purely BMI, contributes to Table 4. Univariable and Multivariable Analysis of All-­Cause Mortality at 2 Years in Patients With Morbid Obesity (n=910) Univariable Analysis Multivariable Analysis HR (95% CI) P Value HR (95% CI) P Value COPD 1.44 (1.01–­2.07) 0.047 eGFR per 10 mL/min per 1.73m2 1.15 (1.06–­1.25) 0.001 1.16 (1.07–­1.26) <0.001 decrease Baseline hemoglobin* 1.32 (1.06–­1.64) 0.013 Major vascular complication 2.01 (1.51–­2.68) <0.001 Life-­threatening bleeding or major 2.66 (1.69–­4.18) <0.001 bleeding Nontransfemoral access 1.70 (1.37–­2.10) <0.001 1.82 (1.19–­2.79) 0.006 General anesthesia 1.40 (0.98–­1.98) 0.062 AKI stage II-­III 3.94 (2.16–­7.17) <0.001 Periprocedure stroke 4.92 (2.30–­10.56) <0.001 AKI indicates acute kidney injury; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; and HR, hazard ratio. *For every 2-­g decrease. J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.01905110 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR A B 95.3% 95.3% Freedom from all-cause mortality (%) 100 94.5% 100 90.4% Freedom from CV mortality (%) 89.4% 83.3% 88.0% 75 74.3% 75 50 50 Log rank, p=0.026 Log rank, p=0.013 25 25 VAT:SAT <1 VAT:SAT<1 VAT:SAT >=1 VAT:SAT>=1 0 0 0 6 12 18 24 0 6 12 18 24 Time (months) Time (months) Number at risk Number at risk VAT:SAT <1 137 99 80 53 39 VAT:SAT<1 137 99 80 53 39 VAT:SAT >=1 82 54 43 28 22 VAT:SAT>=1 82 54 43 28 22 Figure 4. Kaplan-­Meier graphs demonstrating 2-­y all cause (A) and cardiovascular (B) mortality for patients with visceral to subcutaneous adipose tissue (VAT:SAT) ratio <1 and ≥1. CV indicates cardiovascular; SAT, subcutaneous adipose tissue; and VAT, visceral adipose tissue. the patient’s overall risk. Although VAT:SAT ratio did endocrine dysfunction, and implementation of dedi- not correlate with periprocedural events, its predic- cated weight-­loss strategies.43 tive power for midterm outcomes serves to provide an opportunity to initiate interventions after TAVR, Limitations aimed at reducing overall risk, such as a cardiac Although our study represents a large multicenter rehabilitation program. Whether this could improve analysis of the impact of morbid obesity on out- mid-­and longer-­ term outcomes should be a new comes following TAVR, a number of limitations must focus for investigation. be acknowledged. First, this is a retrospective analy- Epicardial adipose tissue, another form of ecto- sis of prospectively collected data and, as such, has pic visceral fat deposition, also predicted all-­cause limitations inherent to this study design. Second, mortality at 2 years. Epicardial fat is thought to have although propensity matching attempts to com- a local, rather than systemic, endocrine effect25 and pensate for the lack of randomization and minimize has been associated with coronary calcification and baseline differences between groups, a number of significant coronary artery stenosis, although it has differences remained, and the possibility of unidenti- been suggested that this association is stronger in fied confounders cannot be excluded. Third, a num- nonobese populations.40 Several studies have re- ber of TAVR CT scans were unsuitable for analysis ported an association with prevalent cardiovascular mainly due to narrowed field of view, which may have disease and with major cardiovascular events,25,41 impacted on our findings relating to the body com- consistent with our findings. A high VAT:SAT ratio, position analysis performed. Still, this is the largest high-­percentage VAT, and high epicardial fat repre- CT analysis to date in the TAVR field regarding this sent an adverse metabolic phenotype that has in- subgroup of patients who are MO. creasingly become of interest.42 Clinical Applicability CONCLUSIONS From a patient-­ centered perspective, the results of TAVR in patients who are MO is a safe procedure and our study may contribute to better decision making demonstrates similar short-­and midterm outcomes in at the time of planning TAVR. With BMI not predicting comparison with nonobese patients, despite slightly midterm outcomes in our cohort, assessment of a pa- higher vascular complications and lower device suc- tient’s obesity phenotype using the pre-­TAVR CT scan cess rates. Adverse obesity phenotype is an important may be a more useful tool to further risk stratify patients predictor of midterm outcomes and body composition who are MO undergoing TAVR. Furthermore, these pa- analysis performed on baseline pre-­TAVR CT can fur- rameters are easily and quickly derived from the pre-­ ther risk stratify patients beyond baseline BMI. TAVR CT with many commercially available software packages providing semiautomatic body composition analysis. As a VAT:SAT ratio ≥1 identifies worse pa- ARTICLE INFORMATION tient outcomes resulting from cardiac and noncardiac Received October 21, 2020; accepted March 29, 2021. causes, diligent action in modifying associated comor- Affiliations bidities should be taken. These include strictly control- Cardiovascular Institute, Hospital Clínico San Carlos, Instituto de Investigación ling cardiovascular risk factors, monitoring patients for Sanitaria San Carlos, Madrid, Spain (A.M., G.T., M.C., N.G., J.E., L.N.); Quebec J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.01905111 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR Heart and Lung Institute, Laval University, Quebec City, Quebec, Canada 4. Siontis GCM, Overtchouk P, Cahill TJ, Modine T, Prendergast B, Praz (J.R., L.G., E.D.); Cardiology Department, Rangueil University Hospital, F, Pilgrim T, Petrinic T, Nikolakopoulou A, Salanti G, et al. Transcatheter Toulouse, France (F.C., A.M.); The Ochsner Clinical School, Ochsner Medical aortic valve implantation vs. surgical aortic valve replacement for treat- Center, New Orleans, LA (J.D.T.S., N.G.E.); Ferrarotto Hospital, University of ment of symptomatic severe aortic stenosis: an updated meta-­analysis. Catania, Catania, Italy (M.B., R.V., E.C.); Centro de Investigación Biomédica Eur Heart J. 2019;40:3143–­3153. DOI: 10.1093/eurhe​artj/ehz275. en Red Enfermedades Cardiovasculares Cardiology Department, Hospital 5. Wolny R, Maehara A, Liu Y, Zhang Z, Mintz GS, Redfors B, Madhavan Universitario Virgen de la Victoria, Málaga, Spain (E.M., A.J.M.); Division of MV, Smits PC, von Birgelen C, Serruys PW, et al. The obesity paradox Cardiology, St. Michael’s Hospital, Toronto University, Toronto, Ontario, Canada revisited: body mass index and long-­term outcomes after PCI from a (M.A., A.N.C.); Hospital Clínico Universitario de Santiago, CIBERCV, Santiago, large pooled patient-­level database. EuroIntervention. 2020;15:1199–­ Spain (D.L., J.L.); Heart Center Lucerne, Luzerner Kantonsspital, Lucerne, 1208. DOI: 10.4244/EIJ-­D-­19-­0 0467. Switzerland (S.T., F.M.); Hospital Universitario Marques de Valdecilla, IDIVAL, 6. Sharma A, Vallakati A, Einstein AJ, Lavie CJ, Arbab-­Zadeh A, Lopez-­ Santander, Spain (G.V., J.M.d.l.T.H.); Department of Interventional Cardiology Jimenez F, Mukherjee D, Lichstein E. Relationship of body mass index and Angiology, National Institute of Cardiology, Warsaw, Poland (A.P., M.D.); with total mortality, cardiovascular mortality, and myocardial infarction CIBERCV, Instituto de Ciencias del Corazón, Hospital Clínico Universitario de after coronary revascularization: evidence from a meta-­analysis. Mayo Valladolid, Valladolid, Spain (T.S., I.J.A.); Cardiology Unit, Cardio-­Thoracic-­ Clin Proc. 2014;89:1080–­1100. DOI: 10.1016/j.mayocp.2014.04.020. Vascular Department, University Hospital of Bologna, Policlinico S, Orsola–­ 7. Sharma A, Lavie CJ, Elmariah S, Borer JS, Sharma SK, Vemulapalli Malpighi, Bologna, Italy (M.C., F.S.); Cardiology Department, Cardiovascular S, Yerokun BA, Li Z, Matsouaka RA, Marmur JD. Relationship of body Institute, Hospital Clínic, Universidad de Barcelona, Institut d’Investigacions mass index with outcomes after transcatheter aortic valve replacement: Biomèdiques August Pi i Sunyer, Barcelona, Spain (A.R.); Hospital General results from the national cardiovascular data-­STS/ACC TVT registry. Universitari Vall d’Hebron, Barcelona, Spain (V.S., B.G.d.); Department of Mayo Clin Proc. 2020;95:57–­68. DOI: 10.1016/j.mayocp.2019.09.027. Cardiology, Hospital Universitario Puerta de Hierro, Majadahonda, Spain 8. González-­ Ferreiro R, Muñoz-­ García AJ, López-­ Otero D, Avanzas P, (J.F.O.); Cardiology Department, Hospital Universitario de la Princesa, Pascual I, Alonso-­Briales JH, Trillo-­Nouche R, Pun F, Jiménez-­Navarro Instituto de Investigación Sanitaria Princesa (IIS-­IP), Universidad Autónoma MF, Hernández-­García JM, et al. Prognostic value of body mass index de Madrid, CIBER-­CV, Madrid, Spain (F.R., F.A.); and Heart Institute (Instituto in transcatheter aortic valve implantation: a “J”-­ shaped curve. Int J de Coraçāo), Sao Paulo, Brazil (H.B.R., D.C.C.B.). Cardiol. 2017;232:342–­347. DOI: 10.1016/j.ijcard.2016.12.051. 9. Abawi M, Rozemeijer R, Agostoni P, van Jaarsveld RC, van Dongen Acknowledgments CS, Voskuil M, Kraaijeveld AO, Doevendans P, Stella PR. Effect of The authors acknowledge the contributions of Dr Iagoba Cano Alsua, body mass index on clinical outcome and all-­cause mortality in pa- Universidad Complutense Madrid, for his statistical support; Dr Miguel tients undergoing transcatheter aortic valve implantation. Neth Heart J. Angel Rubio, Consultant Endocrinologist, Hospital Clínico San Carlos, for 2017;25:498–­509. DOI: 10.1007/s1247​1-­017-­1003-­2. his direction regarding morbidly obese population–­specific complications 10. Yamamoto M, Mouillet G, Oguri A, Gilard M, Laskar M, Eltchaninoff of interest; and Dr Jennifer Mancio for her advice regarding epicardial fat H, Fajadet J, Iung B, Donzeau-­Gouge P, Leprince P, et al. Effect of analysis. body mass index on 30-­and 365-­day complication and survival rates of transcatheter aortic valve implantation (from the FRench Aortic Sources of Funding National CoreValve and Edwards 2 [FRANCE 2] registry). Am J Cardiol. This study was supported by Fundación Interhospitalaria para la Investigación 2013;112:1932–­1937. Cardiovascular (FIC Foundation) via an unrestricted grant from Abbott. 11. Rossi A, Gaibazzi N, Bellelli G, Nistri S, Cicoira M, Cioffi G, Faden G, Temporelli PL, Faggiano P. Obesity paradox in patients with aortic valve Disclosures stenosis. Protective effect of body mass index independently of age, Dr. Toggweiler is a proctor and consultant for Boston Scientific, Medtronic disease severity, treatment modality and non-­cardiac comorbidities. Int and Biosensors/New Valve Technology, a proctor for Abbott Vascular, a con- J Cardiol. 2014;176:1441–­1443. DOI: 10.1016/j.ijcard.2014.08.037. sultant for Shockwave, Teleflex, Medira, AtHeart and holds equity in Hi-­D 12. Konigstein M, Havakuk O, Arbel Y, Finkelstein A, Ben-­Assa E, Leshem Imaging. Dr. Dabrowski is a proctor and consultant for Boston Scientific, has Rubinow E, Abramowitz Y, Keren G, Banai S. The obesity paradox received speaker’s honoraria from Boston Scientific, Edwards Lifesciences in patients undergoing transcatheter aortic valve implantation. Clin Inc. and Medtronic. Cardiol. 2015;38:76–­81. DOI: 10.1002/clc.22355. Dr Regueiro is a proctor for Abbott. Dr Ribeiro is consultant for Edwards 13. van der Boon RM, Chieffo A, Dumonteil N, Tchetche D, Van Mieghem Lifesciences, Medtronic, and Boston Scientific. Dr Garcia del Blanco is proctor NM, Buchanan GL, Vahdat O, Marcheix B, Serruys PW, Fajadet J, et al. for Edwards Lifesciences. Dr Amat-­Santos is a proctor for Boston Scientific. Effect of body mass index on short-­and long-­term outcomes after tran- Dr Saia has received speaker’s honoraria from Edwards Lifesciences Inc, scatheter aortic valve implantation. Am J Cardiol. 2013;111:231–­236. Medtronic, Abbott Vascular, and Boston Scientific. Dr Nombela-­Franco has DOI: 10.1016/j.amjca​rd.2012.09.022. served as a proctor for Abbott and has received speaker’s honoraria from 14. Gonzalez Ferreiro R, Lopez Otero D, Alvarez Rodriguez L, Iglesias Edwards Lifesciences Inc and Boston Scientific. The remaining authors have Alvarez D, Cid Menendez A, Antunez Muinos PJ, Cid Alvarez no disclosures to report. B, Sanmartin Pena XC, Cruz-­ G onzalez I, Trillo Nouche R, et al. Transcatheter aortic valve replacement in patients with morbid Supplementary Material obesity: procedure outcomes and medium-­ term follow-­ u p. JACC Video S1 Cardiovasc Interv. 2019;12:1192–­1193. Table S1–­S5 15. Vecchie A, Dallegri F, Carbone F, Bonaventura A, Liberale L, Portincasa Figure S1–­S3 P, Fruhbeck G, Montecucco F. Obesity phenotypes and their par- adoxical association with cardiovascular diseases. Eur J Intern Med. 2018;48:6–­17. DOI: 10.1016/j.ejim.2017.10.020. 16. Rodes-­Cabau J. Transcatheter aortic valve implantation: current and REFERENCES future approaches. Nat Rev Cardiol. 2011;9:15–­29. DOI: 10.1038/nrcar​ 1. World Health Organization. Obesity and overweight 2020. Available dio.2011.164. from: https://www.who.int/news-­room/fact-­sheet​s/detai​l/obesi​t y-­and-­ 17. Gastrointestinal surgery for severe obesity. National institutes of overw​eight. Accessed June 05, 2020. health consensus development conference statement. Am J Clin Nutr. 2. Ward ZJ, Bleich SN, Cradock AL, Barrett JL, Giles CM, Flax C, Long 1992;55:615S–­619S. MW, Gortmaker SL. Projected US state-­level prevalence of adult obe- 18. Obesity: preventing and managing the global epidemic. Report of a WHO sity and severe obesity. N Engl J Med. 2019;381:2440–­ 2450. DOI: consultation. World Health Organ Tech Rep Ser. 2000;894:i-­xii, 1-­253. 10.1056/NEJMs​a1909301. 19. Kappetein AP, Head SJ, Généreux P, Piazza N, van Mieghem NM, 3. Ando T, Akintoye E, Trehan N, Telila T, Briasoulis A, Takagi H, Grines Blackstone EH, Brott TG, Cohen DJ, Cutlip DE, van Es GA, et al. CL, Afonso L. Comparison of in-­ hospital outcomes of transcath- Updated standardized endpoint definitions for transcatheter aortic eter aortic valve implantation versus surgical aortic valve replace- valve implantation: the valve academic research consortium-­2 consen- ment in obese (body mass index ≥ 30 kg/m2) patients. Am J Cardiol. sus document. J Am Coll Cardiol. 2012;60:1438–­1454. DOI: 10.1016/j. 2017;120:1858–­1862. jacc.2012.09.001. J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.01905112 McInerney et al Morbid Obesity and Obesity Phenotype in TAVR 20. Hahn RT, Leipsic J, Douglas PS, Jaber WA, Weissman NJ, Pibarot 32. Kotronias RA, Scarsini R, De Maria GL, Rajasundaram S, Sayeed R, P, Blanke P, Oh JK. Comprehensive echocardiographic assessment Krasopoulos G, Grebenik C, Keiralla A, Newton JD, Banning AP, et of normal transcatheter valve function. JACC Cardiovasc Imaging. al. Ultrasound guided vascular access site management and left ven- 2019;12:25–­34. DOI: 10.1016/j.jcmg.2018.04.010. tricular pacing are associated with improved outcomes in contempo- 21. Mok M, Allende R, Leipsic J, Altisent J, del Trigo M, Campelo-­Parada rary transcatheter aortic valve replacement: insights from the OxTAVI F, DeLarochellière R, Dumont E, Doyle D, Côté M, et al. Prognostic registry. Catheter Cardiovasc Interv. 2020;96:432–­439. DOI: 10.1002/ value of fat mass and skeletal muscle mass determined by computed ccd.28578. tomography in patients who underwent transcatheter aortic valve 33. Britton KA, Massaro JM, Murabito JM, Kreger BE, Hoffmann U, Fox implantation. Am J Cardiol. 2016;117:828–­8 33. DOI: 10.1016/j.amjca​ CS. Body fat distribution, incident cardiovascular disease, cancer, and rd.2015.12.015. all-­cause mortality. J Am Coll Cardiol. 2013;62:921–­925. DOI: 10.1016/j. 22. Faron A, Luetkens JA, Schmeel FC, Kuetting DLR, Thomas D, Sprinkart jacc.2013.06.027. AM. Quantification of fat and skeletal muscle tissue at abdominal com- 34. Lee JJ, Pedley A, Hoffmann U, Massaro JM, Fox CS. Association of puted tomography: associations between single-­slice measurements changes in abdominal fat quantity and quality with incident cardiovas- and total compartment volumes. Abdom Radiol (NY). 2019;44:1907–­ cular disease risk factors. J Am Coll Cardiol. 2016;68:1509–­1521. 1916. DOI: 10.1007/s0026​1-­019-­01912​-­9. 35. Mancio J, Fonseca P, Figueiredo B, Ferreira W, Carvalho M, Ferreira 23. Mourtzakis M, Prado CM, Lieffers JR, Reiman T, McCargar LJ, Baracos N, Braga P, Rodrigues A, Barros A, Falcao-­Pires I, et al. Association VE. A practical and precise approach to quantification of body compo- of body mass index and visceral fat with aortic valve calcification and sition in cancer patients using computed tomography images acquired mortality after transcatheter aortic valve replacement: the obesity par- during routine care. Appl Physiol Nutr Metab. 2008;33:997–­1006. DOI: adox in severe aortic stenosis. Diabetol Metab Syndr. 2017;9:86. DOI: 10.1139/H08-­075. 10.1186/s1309​8-­017-­0285-­2. 24. Capoulade R, Larose É, Mathieu P, Clavel MA, Dahou A, Arsenault M, 36. Okuno T, Koseki K, Nakanishi T, Ninomiya K, Tomii D, Tanaka T, Bédard É, Larue-­Grondin S, Le Ven F, Dumesnil JG, et al. Visceral adi- Sato Y, Osanai A, Sato K, Koike H, et al. Prognostic impact of com- posity and left ventricular mass and function in patients with aortic ste- puted tomography-­derived abdominal fat area on transcatheter aor- nosis: the PROGRESSA study. Can J Cardiol. 2014;30:1080–­1087. DOI: tic valve implantation. Circ J. 2018;82:3082–­3089. DOI: 10.1253/circj. 10.1016/j.cjca.2014.02.007. CJ-­18-­0709. 25. Marwan M, Achenbach S. Quantification of epicardial fat by computed 37. Mechanick JI, Farkouh ME, Newman JD, Garvey WT. Cardiometabolic-­ tomography: why, when and how? J Cardiovasc Comput Tomogr. based chronic disease, adiposity and dysglycemia drivers: JACC state-­ 2013;7:3–­10. DOI: 10.1016/j.jcct.2013.01.002. of-­the-­art review. J Am Coll Cardiol. 2020;75:525–­538. DOI: 10.1016/j. 26. Abramowitz Y, Chakravarty T, Jilaihawi H, Cox J, Sharma RP, Mangat jacc.2019.11.044. G, Nakamura M, Cheng W, Makkar RR. Impact of body mass index on 38. Rodriguez-­Granillo GA, Reynoso E, Capunay C, Carpio J, Carrascosa the outcomes following transcatheter aortic valve implantation. Catheter P. Pericardial and visceral, but not total body fat, are related to global Cardiovasc Interv. 2016;88:127–­134. DOI: 10.1002/ccd.26394. coronary and extra-­ coronary atherosclerotic plaque burden. Int J 27. Koifman E, Kiramijyan S, Negi SI, Didier R, Escarcega RO, Minha S, Cardiol. 2018;260:204–­210. DOI: 10.1016/j.ijcard.2018.01.106. Gai J, Torguson R, Okubagzi P, Ben-­Dor I, et al. Body mass index as- 39. Ladeiras-­ Lopes R, Sampaio F, Bettencourt N, Fontes-­ Carvalho R, sociation with survival in severe aortic stenosis patients undergoing Ferreira N, Leite-­Moreira A, Gama V. The ratio between visceral and transcatheter aortic valve replacement. Catheter Cardiovasc Interv. subcutaneous abdominal fat assessed by computed tomography is an 2016;88:118–­124. DOI: 10.1002/ccd.26377. independent predictor of mortality and cardiac events. Rev Esp Cardiol 28. Dayan V, Vignolo G, Soca G, Paganini JJ, Brusich D, Pibarot P. (Engl Ed). 2017;70:331–­337. Predictors and outcomes of prosthesis-­ patient mismatch after 40. Gorter PM, de Vos AM, van der Graaf Y, Stella PR, Doevendans PA, Meijs aortic valve replacement. JACC Cardiovasc Imaging. 2016;9: MF, Prokop M, Visseren FL. Relation of epicardial and pericoronary fat 924–­933. to coronary atherosclerosis and coronary artery calcium in patients un- 29. Mohty D, Dumesnil JG, Echahidi N, Mathieu P, Dagenais F, Voisine P, dergoing coronary angiography. Am J Cardiol. 2008;102:380–­385. DOI: Pibarot P. Impact of prosthesis-­patient mismatch on long-­term survival 10.1016/j.amjca​rd.2008.04.002. after aortic valve replacement: influence of age, obesity, and left ven- 41. Mahabadi AA, Massaro JM, Rosito GA, Levy D, Murabito JM, Wolf PA, tricular dysfunction. J Am Coll Cardiol. 2009;53:39–­47. DOI: 10.1016/j. O’Donnell CJ, Fox CS, Hoffmann U. Association of pericardial fat, in- jacc.2008.09.022. trathoracic fat, and visceral abdominal fat with cardiovascular disease 30. Seoudy H, Gussefeld N, Frank J, Freitag-­ Wolf S, Lutter G, Eden burden: the Framingham Heart Study. Eur Heart J. 2009;30:850–­856. M, Rangrez AY, Kuhn C, Frey N, Frank D. Incidence and impact of DOI: 10.1093/eurhe​artj/ehn573. prosthesis-­patient mismatch following transcatheter aortic valve im- 42. Piche ME, Poirier P, Lemieux I, Despres JP. Overview of epidemiology plantation. Clin Res Cardiol. 2019;108:660–­668. DOI: 10.1007/s0039​ and contribution of obesity and body fat distribution to cardiovascu- 2-­018-­1394-­0. lar disease: an update. Prog Cardiovasc Dis. 2018;61:103–­113. DOI: 31. Hibbert B, Simard T, Wilson KR, Hawken S, Wells GA, Ramirez FD, Le 10.1016/j.pcad.2018.06.004. May MR, So DY, Glover CA, Froeschl M, et al. Transradial versus trans- 43. Bray GA, Heisel WE, Afshin A, Jensen MD, Dietz WH, Long M, Kushner femoral artery approach for coronary angiography and percutaneous RF, Daniels SR, Wadden TA, Tsai AG, et al. The science of obesity coronary intervention in the extremely obese. JACC Cardiovasc Interv. anagement: an endocrine society scientific statement. Endocr Rev. 2012;5:819–­826. DOI: 10.1016/j.jcin.2012.04.009. 2018;39:79–­132. DOI: 10.1210/er.2017-­0 0253 J Am Heart Assoc. 2021;10:e019051. DOI: 10.1161/JAHA.120.01905113 SUPPLEMENTAL MATERIAL Table S1. Procedural aspects in unmatched non-obese and morbidly obese cohorts. BMI 18.5-29.9 BMI >35 P (n= 2,264) (n= 910) Procedural urgency Urgent/Emergent 180 (9.7%) 67 (8.2%) 0.224 Access site Transfemoral 1994 (88.1%) 784 (86.2%) 0.139 Non-transfemoral access 270 (11.9%) 126 (13.9%) 0.139 Method of transfemoral access*† Percutaneous with closure 1626 (93.9%) 678 (89.8%) <0.001 device Surgical cut down 105 (6.1%) 77 (10.2%) <0.001 Prosthesis type BEV 899 (39.7%) 432 (47.5%) <0.001 SEV 1362 (60.2%) 469 (51.5%) <0.001 Edwards Sapien & XT & S3 899 (39.7%) 432 (47.5%) <0.001 Medtronic Corevalve, Evolut R, Evolut 1065 (47.0%) 403 (44.3%) 0.159 Pro Other (portico, accurate neo, 300 (13.2%) 75 (8.2%) <0.001 other) Prosthesis size 20 – 23 mm 525 (23.5%) 207 (22.9%) 0.782 25 – 27 mm 1005 (45.0%) 394 (44.1) 0.650 29 – 34 mm 702 (31.5%) 293 (32.7%) 0.485 Other procedural aspects General anesthesia 958 (42.3%) 350 (38.5%) 0.048 Prior balloon valvuloplasty 1367 (65.9%) 498 (60.1%) 0.003 Balloon post-dilatation 511 (23.6%) 112 (12.5%) <0.001 *= One center excluded that practices only ‘cut down’ technique for femoral access †= femoral access only BEV: Balloon expandable valve, SEV: Self expanding valve Table S2. Clinical end points and echocardiographic data post procedure for unmatched non-obese and morbidly obese cohorts. BMI 18.5-29.9 BMI >35 Clinical endpoints p (n= 2,264) (n= 910) Mortality In-hospital mortality 82 (3.6%) 35 (3.9%) 0.762 In-hospital or 30-day mortality 86 (3.8%) 43 (4.7%) 0.232 Vascular complications Major 103 (4.6%) 60 (6.6%) 0.019 Minor 183 (10.3%) 67 (8.7%) 0.196 Vascular complications femoral access only Major 96 (4.8%) 55 (7.0%) 0.022 Minor 182 (11.8%) 62 (9.2%) 0.072 Major vascular complications femoral access only by closure method* Percutaneous closure device 71 (4.4%) 44 (6.5%) 0.034 Surgical cut-down technique 10 (9.5%) 11 (14.3%) 0.321 Bleeding Life-threatening bleeding 69 (3.1%) 23 (2.6%) 0.454 Major bleeding 134 (5.9%) 51 (5.6%) 0.737 Life-threatening and major 203 (9.0%) 74 (8.1%) 0.451 Minor bleeding 217 (9.6%) 61 (6.9%) 0.015 AKI Stage I 221 (14.4%) 118 (14.5%) 0.920 Stage II and III 55 (3.6%) 31 (3.8%) 0.769 Any stage 276 (18.0%) 149 (18.4%) 0.814 Coronary occlusion 11 (0.5%) 5 (0.5%) 0.819† Peri-Procedural Stroke 39 (1.7%) 14 (1.5%) 0.714 Hospital acquired pneumonia 39 (1.9%) 11 (1.2%) 0.236 New permanent pacemaker 228 (11.2%) 121 (14.7%) 0.010 implantation Length of hospital stay, days 6 [5-9] 5 [3-8] <0.001 Echocardiogram parameters within 30-day post-TAVR Moderate-Severe Post TAVR 104 (4.7%) 21 (2.4%) 0.004 AR Postprocedural mean aortic 8 [6-11] 10 [7-14] <0.001 valve gradient (mmHg) Severe patient-prosthesis 20 (1.1%) 27 (3.5%) <0.001 mismatch Device Success 2031 (89.7%) 759 (83.4%) <0.001 Echocardiogram parameters at 1-year post TAVR Mean aortic valve gradient 8 [5.9-10.8] 10 [7-15] <0.001 (mmHg) Values are expressed as n (%) or median [IQR] * One center excluded that practices only ‘cut down’ technique for femoral access † Fischers exact test used AR: aortic regurgitation, TAVR: transcatheter aortic valve replacement. Table S3. Univariable and multivariable analysis of all-cause mortality at 2 years in the whole cohort (non-obese and morbidly obese patients n=3174). Univariable Multivariable analysis p value analysis p value HR (95% CI) HR (95% CI) Morbid obesity 1.01 (0.83-1.25) 0.893 BMI 0.99 (0.99-1.01) 0.826 BSA 1.16 (0.81-1.66) 0.430 Diabetes 1.24 (1.03-1.49) 0.026 Hypertension 1.32 (1.00-1.74) 0.049 Smoking 1.25 (1.01-1.55) 0.033 COPD 1.44 (1.18-1.75) <0.001 1.38 (1.10-1.74) 0.006 Severe pulmonary hypertension 1.49 (1.17-1.90) 0.001 (>55mmHg) Peripheral vascular disease 1.59 (1.26-1.99) <0.001 Pre-existing atrial fibrillation 1.45 (1.20-1.74) <0.001 eGFR <30 2.24 (1.74-2.88) <0.001 Baseline Haemoglobin * 1.26 (1.13-1.41) <0.001 Moderate-Severe MR 1.29 (1.03-1.61) 0.024 1.35 (1.05-1.75) 0.022 Urgent/Emergent procedure 1.73 (1.29-2.33) <0.001 Non-transfemoral access 1.66 (1.32-2.09) <0.001 1.51 (1.16-1.97) 0.002 Conversion to surgery 6.06 (2.87-12.79) <0.001 Hospital acquired pneumonia 3.85 (2.56-5.82) <0.001 Major vascular complications 1.97 (1.43-2.72) <0.001 Life threatening or major bleeding 2.31 (1.80-2.96) <0.001 1.86 (1.39-2.48) <0.001 Blood transfusion 1.69 (1.39-2.07) <0.001 Periprocedural CVA 3.49 (2.27-5.35) <0.001 New onset atrial fibrillation 1.33 (1.00-1.766) 0.051 Post procedure moderate-severe AR 1.79 (1.13-2.84) 0.013 In-hospital days 1.01 (1.01-1.02) <0.001 AKI stage II-III 4.65 (3.31-6.54) <0.001 3.88 (2.72-5.53) <0.001 AKI: acute kidney injury, AR: aortic regurgitation, BMI: body mass index, BSA: body surface area, COPD: Chronic obstructive pulmonary disease, CVA: cerebrovascular accident, eGFR: estimated glomerular filtration rate, MR: mitral regurgitation. *For every 2gram decrease Table S4. Summary of body composition analysis. Body composition component Mean (SD) SAT area cm2 357.9 (118.3) iSAT area cm2/m2 172.4 (56.3) VAT area cm2 311.3 (128.6) 2 2 iVAT area cm /m 146.0 (53.0) VAT:SAT 1.03 (0.77) Percentage Visceral adipose tissue 46 (13) IMAT cm2 35.2 (19.0) iIMAT (indexed intramuscular adipose tissue) cm2/m2 16.6 (8.7) Percentage fatty muscle (%) 22.0 (10.0) IMAT:SMA 0.30 (0.19) SMA cm2 127.3 (33.7) 2 2 iSMA area cm /m 60.2 (13.5) Sarcopenic obesity* 7.84% EAT volume cm3 98.3 (50.4) iEAT cm3/m2 47.0 (23.1) Values are presented as mean and standard deviation (SD). Indexed values are indexed to body surface area (BSA) *Sarcopenic obesity defined as height indexed skeletal muscle area (hiSMA) ≤ 38.5cm2/m2 for females and ≤ 52.5cm2/m2 for males EAT: epicardial adipose tissue, iEAT: indexed epicardial adipose tissue, IAT: Intramuscular adipose tissue, iIMAT: indexed intramuscular adipose tissue, IMAT:SMA ratio of intramuscular fat area to skeletal muscle area, SAT: subcutaneous adipose tissue, iSAT: indexed subcutaneous fat, SMA: skeletal muscle area, iSMA: indexed skeletal muscle area, VAT: Visceral adipose tissue, iVAT: indexed visceral adipose tissue, VAT:SAT: ratio of visceral adipose tissue area to subcutaneous adipose tissue area. Table S5. Baseline characteristics in morbidly obese cohort according to VAT:SAT ratio. VAT:SAT ratio <1 VAT:SAT ratio 1 p (n= 137) (n= 82) Age, years 77.4 (7.34) 77.3 (6.7) 0.892 Female sex 118 (86.1%) 18 (23.2%) <0.001 Body mass index, 39.0 (3.9) 38.5 (3.03) 0.323 kg/m2 Diabetes mellitus 75 (54.74%) 53 (64.63%) 0.151 -Insulin use 27 (37.0%) 29 (56.9%) 0.029 Hypertension 123 (89.8%) 78 (95.1%) 0.164 Hyperlipidaemia 102 (74.5%) 58 (73.4%) 0.867 Baseline creatinine 1.09 (0.46) 1.40 (0.80) <0.001 (mg/dL) eGFR <30 8 (5.8%) 9 (11.0%) 0.196* ml/min/1.73m2 Coronary artery disease 58 (42.3%) 46 (56.1%) 0.048 Pre valve surgery 9 (6.6%) 4 (4.9%) 0.608 Valve in Valve TAVR 6 (5.2%) 2 (2.7%) 0.468* Atrial fibrillation 45 (32.9%) 26 (31.7%) 0.862 Previous pacemaker 13 (9.5%) 10 (12.2%) 0.649 COPD 35 (25.6%) 28 (34.1%) 0.174 Previous cerebrovascular 11 (8.0%) 8 (9.8%) 0.660 accident Peripheral vascular 16 (11.7%) 15 (18.3%) 0.174 disease Baseline hemoglobin 11.9 (1.6) 12.2 (1.8) 0.297 (g/dL) NT-Pro BNP (median, 877.45 [340 -1897] 887.5 [273-1677.5] 0.981 IQR) Logistic EuroSCORE 11.63 [7.05-17.02] 11.2 [8.12-17.0] 0.741 [median IQR] EuroSCORE II 2.97 [1.94-5.52] 4.56 [2.42-6.96] 0.026 [median IQR] STS [median IQR] 4.05 [2.9-6.71] 4.1 [3.01-6.40] 0.968 Moderate or severe 57 (47.9%) 36 (54.6%) 0.386 PHT Values are expressed as n (%) or median [IQR] *Fischer exact test used eGFR: estimated glomerular filtration rate, PHT: pulmonary hypertension, TAVR: transcatheter aortic valve replacement A B C D E F Figure S1. Typical cardiac CT, with contrast, demonstrating the fibrous pericardium (A, white arrows) and subsequent segmentation of epicardial adipose tissue (EAT) within the fibrous pericardium. Using a software package, EAT was segmented based on a Hounsfields unit thresholds of -190 to -30 from the bifurcation of the pulmonary artery (B) continuing to the diaphragm. Images C-E represent examples of slices taken from the mid atrial level, mid- ventricular 4-chamber and lower ventricular 4 chamber view respectively. Epicardial adipose tissue was manually redefined every 3 slices in order to correct contours and avoid inclusion of paracardial adipose tissue (outside the pericardial sac). The software calculated the EAT volume (cm3) by summing the EAT area in each slice and taking into account slice thickness and intersection gap. Image F represents a 3D reconstruction of the EAT in this patient (a female with BMI 35.4kg/m2 and EAT volume of 108.6 cm3 and iEAT of 56.77 cm3/m2). Figure S2. Density plots showing the density index pre- (A) and post- (B) propensity- score matching. Figure S3. Kaplan Meier graph demonstrating 2-year all-cause (A) and cardiovascular (B) mortality and 2-year all-cause (C) and cardiovascular (D) readmissions for non- obese and morbidly obese groups in the unmatched cohort. Supplemental Video Legend: Video S1. 3D reconstruction of epicardial fat segmentation. The video demonstrates assessment of epicardial fat from the bifurcation of the pulmonary artery to the diaphragm. Best viewed with Windows Media Player.