Limited data have suggested that an “obesity paradox” exists for mortality and cardiovascular outcomes in patients undergoing coronary artery bypass grafting. Much less is known about the role of the preoperative body mass index (BMI) in patients undergoing valve surgery. We evaluated 2,640 consecutive patients who underwent valve surgery between April 2004 and March 2011. The patients were classified by the World Health Organization standards as “underweight” (BMI 11.5 to 18.4 kg/m 2 , n = 61), “normal weight” (BMI 18.5 to 24.9 kg/m 2 , n = 865), “overweight” (BMI 25 to 29.9 kg/m 2 , n = 1,020), and “obese” (BMI 30 to 60.5 kg/m 2 , n = 694). Mortality was ascertained using the Social Security Death Index. Hazard ratios (HRs), adjusted for known preoperative risk factors, were obtained using Cox regression models. The mean follow-up was 31.9 ± 20.5 months. The long-term mortality rate was 1.21, 0.52, 0.32, and 0.44 per 10 years of person-time for underweight, normal, overweight, and obese patients, respectively. Compared to the normal BMI category, overweight patients (adjusted HR 0.60, 95% confidence interval 0.46 to 0.79, p <0.001) and obese patients (adjusted HR 0.67, 95% confidence interval 0.50 to 0.91, p = 0.009) were at a lower hazard of long-term all-cause mortality. Underweight patients remained at a greater adjusted risk of long-term mortality than normal weight patients (adjusted HR 1.69, 95% confidence interval 1.01 to 2.85, p = 0.048). Similar patterns of mortality outcomes were noted in the subset of patients undergoing isolated valve surgery. In conclusion, overweight and obese patients had greater survival after valve surgery than patients with a normal BMI. Very lean patients undergoing valve surgery are at a greater hazard for mortality and might require more rigorous preoperative candidate screening and closer postoperative monitoring.
Few studies have examined the association of body mass index (BMI) with outcomes in patients undergoing valvular heart surgery, especially those undergoing isolated valvular procedures. The evaluation of this population could provide insight into the influence of body mass as it relates to acute perturbations that occur in the perioperative period of cardiac surgery, independent of underlying coronary artery disease. We sought to determine whether differences in the preoperative BMI are associated with the postoperative course and late survival of patients undergoing valvular heart surgery, with a specific focus on patients undergoing isolated valvular procedures.
Methods
We retrospectively analyzed the data from 2,640 consecutive patients who underwent valvular heart surgery at Northwestern Memorial Hospital between April 1, 2004 and March 31, 2011. Patients undergoing concomitant coronary artery bypass grafting or atrial fibrillation ablation were included. All co-morbidities and outcomes were defined and classified according to the standard definitions established by the Society of Thoracic Surgeons (STS) national database (available from: http://www.sts.org ). The patients were stratified by BMI into 4 categories, as defined by the World Health Organization : underweight (BMI ≥11.5 to <18.5 kg/m 2 ), normal weight (BMI ≥18.5 to <25 kg/m 2 ), overweight (BMI ≥25 to <30 kg/m 2 ), and obese (BMI ≥30 to <60.5 kg/m 2 ). Data on postoperative events and late survival were available for all patients included in the study cohort. The baseline demographic, clinical, and surgical characteristics were collected from the Bluhm Cardiovascular Institute’s Clinical Trials Units Cardiovascular Research Database and medical chart review, as needed. This database has been approved by the institutional review board at Northwestern University (project no. STU00012288). Any subjects refusing participation in the project were not included in the analysis.
The primary end point was all-cause mortality, determined from the Social Security Death Index, and assessed up to March 2011. The secondary outcomes were in-hospital mortality, 30-day mortality, postoperative complications, length of stay (LOS), and discharge medications. The cause of death was ascertained by chart review for all patients who died during hospitalization for index valve surgery. The LOS included the total number of days spent in the hospital and the number of hours in the intensive care unit.
All preoperative, intraoperative, and postoperative variables were compared across the 4 BMI groups. The chi-square test and Fisher’s exact test were used to compare the categorical variables across the 4 groups. Continuous variables were analyzed using analysis of variance and Student-Newman-Keuls tests for post hoc comparisons. Unadjusted all-cause long-term mortality was summarized using Kaplan-Meier curves, and the BMI groups were compared using the log-rank test. Weighted log-rank tests with Fleming-Harrington weights were used to assign differential weights to early, middle, or late mortality differences. Cox proportional hazards regression was used to compare risk factor-adjusted all-cause long-term mortality among the BMI groups. The predictive ability of a Cox regression model was summarized using the index of concordance and implemented using the SAS macro, % survcstd.
We also analyzed the association of BMI as a continuous variable to the risk of all-cause long-term mortality using a Cox regression model with additive smoothing splines. For software implementation, we used the SAS macro, % PSPLINET (available from: http://biostat.mc.vanderbilt.edu/wiki/Main/SasMacros ). The risk-adjustment models included 28 prespecified variables known to influence postoperative outcomes. The Ambler score is an indexed measure of disease severity and validated in this population; therefore, we used it to control for confounding by indication. The European System for Cardiac Operative Risk Evaluation (euroSCORE) and the STS score are 2 risk stratification models that incorporate a number of demographic and clinical variables in the context of the specific cardiac operation to estimate operative morbidity and mortality.
Prespecified subset analyses were performed in patients undergoing isolated valvular procedures (single and multiple, without concomitant coronary artery bypass grafting or atrial fibrillation ablation) and in the obese patient subgroup. In the latter, 3 separate groups were created using the World Health Organization criteria : obese class I (BMI ≥30.0 to <35 kg/m 2 ), obese class II (BMI ≥35.0 to <40 kg/m 2 ), and obese class III (BMI ≥40.0 to 60.5 kg/m 2 ). Late survival was assessed separately in each of these subsets by BMI class. All statistical analyses were performed using SAS, version 9.2, statistical software (SAS Institute, Cary, North Carolina). No adjustments for multiplicity were made. All tests were 2-sided, and statistical significance was declared when p <0.05.
Results
In the present 7-year surgical experience, 61 patients (2.3%) were classified as underweight (BMI ≥11.5 to <18.5 kg/m 2 ), 865 (32.8%) as normal weight (BMI ≥18.5 to <25 kg/m 2 ), 1,020 (38.6%) as overweight (BMI ≥25 to <30 kg/m 2 ), and 694 (26.3%) as obese (BMI ≥30 to <60.5 kg/m 2 ). Complete demographic, clinical, and surgical characteristics stratified by BMI classification are listed in Table 1 . Data were missing for <2.5% of all included covariates. Patients with higher BMI classes were significantly more likely to have coronary artery disease (p = 0.04), a family history of coronary artery disease (p = 0.019), and cardiac co-morbidities, such as renal impairment, hyperlipidemia, hypertension, and diabetes (all p <0.001). In contrast, underweight patients were more likely to be women (p <0.001) and to have active infectious endocarditis (p = 0.0009), chronic obstructive pulmonary disease (p <0.001), and pulmonary hypertension (p <0.001). Underweight patients had a mean Ambler score 3 to 4 times greater than the score of those in other BMI classes (p <0.001) and consistently had the highest log euroSCOREs and STS scores. The prevalence of worse function classes (p <0.001) and elective status (p = 0.06) were significantly different across the BMI classes, with no discernible trend. Patients in the higher BMI classes were more likely to undergo aortic valve procedures (p <0.001), and patients in the underweight group were more likely to undergo mitral and tricuspid valve surgery (p <0.001). In addition, obese and overweight patients were more likely to be prescribed perioperative angiotensin-converting enzyme inhibitors, aspirin, lipid-lowering medications, and β blockers. In contrast, warfarin was prescribed significantly more often for underweight patients. The baseline profiles of obese class I, II, and III patients are listed separately in Table 2 .
| Variable | BMI (kg/m 2 ) | p Value | |||
|---|---|---|---|---|---|
| 11.5–18.4 | 18.5–24.9 | 25–29.9 | 30–60.5 | ||
| (n = 61) | (n = 865) | (n = 1,020) | (n = 694) | ||
| Women | 46 (75%) | 372 (43%) | 309 (30%) | 288 (41%) | <0.001 A,B,C,D |
| Age (years) | 62.0 ± 17.2 | 63.7 ± 15.9 | 63.2 ± 14.8 | 62.7 ± 12.9 | 0.57 |
| White | 42 (75%) | 664 (83%) | 775 (82%) | 528 (82%) | 0.54 |
| Current smoker | 3 (5%) | 33 (4%) | 33 (3%) | 25 (4%) | 0.85 |
| Diabetes mellitus | 6 (10%) | 73 (8%) | 137 (13%) | 196 (28%) | <0.001 C,D,E |
| Hyperlipidemia ⁎ | 18 (30%) | 380 (44%) | 547 (54%) | 430 (62%) | <0.001 A,B,C,D,E,F |
| Renal failure | 4 (7%) | 38 (4%) | 41 (4%) | 43 (6%) | 0.17 |
| Dialysis | 3 (5%) | 17 (2%) | 15 (1%) | 17 (2%) | 0.18 |
| Hypertension † | 24 (39%) | 436 (50%) | 616 (60%) | 508 (73%) | <0.001 B,C,D,E |
| Infective endocarditis | 7 (11%) | 81 (9%) | 81 (8%) | 35 (5%) | 0.009 C,E,F |
| Chronic obstructive pulmonary disease | 16 (26%) | 101 (12%) | 111 (11%) | 106 (15%) | <0.001 A,B,C,E |
| Peripheral vascular disease | 3 (5%) | 50 (6%) | 67 (7%) | 54 (8%) | 0.42 |
| Cerebrovascular disease | 8 (13%) | 113 (13%) | 107 (10%) | 95 (14%) | 0.18 |
| Cerebrovascular accident | 7 (11%) | 94 (11%) | 90 (9%) | 79 (11%) | 0.3 |
| Previous myocardial infarction | 6 (10%) | 85 (10%) | 100 (10%) | 79 (11%) | 0.71 |
| Heart failure (by history) ‡ | 24 (39%) | 283 (33%) | 333 (33%) | 253 (36%) | 0.25 |
| New York Heart Association class III-IV | 25 (41%) | 285 (33%) | 342 (34%) | 299 (43%) | <0.001 E,F |
| Atrial fibrillation (by history) ‡ | 25 (41%) | 257 (30%) | 288 (28%) | 221 (32%) | 0.1 |
| Previous cardiovascular intervention | 20 (33%) | 258 (30%) | 293 (29%) | 228 (33%) | 0.31 |
| Angina pectoris | 7 (11%) | 102 (12%) | 156 (15%) | 108 (16%) | 0.09 |
| Family history of coronary artery disease § | 6 (10%) | 123 (14%) | 140 (14%) | 129 (19%) | 0.02 E,F |
| Coronary artery disease ∥ | 17 (28%) | 298 (36%) | 393 (39%) | 286 (42%) | 0.04 C,E |
| Repeat sternotomy | 12 (20%) | 166 (19%) | 177 (17%) | 118 (17%) | 0.64 |
| Elective surgery | 48 (79%) | 727 (84%) | 890 (87%) | 581 (84%) | 0.06 |
| Urgent surgery | 11 (18%) | 123 (14%) | 114 (11%) | 104 (15%) | 0.06 |
| Preoperative creatinine (mg/dl) | <0.001 B,C,D,E | ||||
| Median | 0.9 | 1.0 | 1.0 | 1.0 | |
| Interquartile range | 0.7–1.2 | 0.8–1.2 | 0.9–1.2 | >0.9–1.3 | |
| Ejection fraction (%) | 0.15 | ||||
| Median | 55 | 60 | 60 | 60 | |
| Interquartile range | 40–60 | 50–65 | 50–65 | 50–63 | |
| Perfusion time (min) | 0.76 | ||||
| Median | 119 | 113 | 113 | 117 | |
| Interquartile range | 90–166 | 83–156 | 82–154 | 84–156 | |
| Cross-clamp time (min) | 0.43 | ||||
| Median | 91 | 88 | 89.5 | 93 | |
| Interquartile range | 62–123 | 68–125 | 68–121 | 69–126 | |
| Underlying valve disease | |||||
| Aortic stenosis | 20 (34%) | 275 (34%) | 388 (40%) | 328 (49%) | <0.001 C,D,E,F |
| Aortic regurgitation ≥2+ | 17 (28%) | 232 (28%) | 249 (25%) | 164 (25%) | 0.51 |
| Mitral stenosis | 10 (18%) | 79 (10%) | 78 (8%) | 65 (10%) | 0.11 |
| Mitral regurgitation ≥2+ | 40 (66%) | 496 (59%) | 469 (47%) | 268 (39%) | <0.001 B,C,D,E |
| Tricuspid regurgitation ≥2+ | 25 (41%) | 205 (24%) | 185 (19%) | 105 (16%) | <0.001 A,B,C,D |
| Surgery type | |||||
| Isolated valve | 24 (39%) | 314 (36%) | 343 (34%) | 224 (32%) | 0.3 |
| Aortic | 8 (13%) | 110 (13%) | 160 (16%) | 129 (19%) | 0.015 E |
| Mitral | 7 (12%) | 155 (18%) | 135 (13%) | 61 (9%) | <0.001 D,E,F |
| All aortic valve | 38 (62%) | 445 (51%) | 612 (60%) | 468 (67%) | <0.001 |
| Repair | 12 (32%) | 151 (34%) | 215 (35%) | 147 (31%) | 0.63 |
| Replacement | 26 (68%) | 294 (66%) | 397 (65%) | 321 (69%) | 0.63 |
| All mitral valve | 37 (61%) | 497 (57%) | 474 (46%) | 281 (40%) | <0.001 B,C,D,E |
| Repair | 22 (59%) | 368 (74%) | 336 (71%) | 174 (62%) | 0.002 E,F |
| Replacement | 15 (41%) | 129 (26%) | 138 (29%) | 107 (38%) | 0.002 E,F |
| All tricuspid valve | 18 (30%) | 145 (17%) | 138 (14%) | 78 (11%) | <0.001 A,B,C,E |
| Repair | 17 (94%) | 135 (93%) | 124 (90%) | 73 (94%) | 0.68 |
| Replacement | 1 (6%) | 10 (7%) | 14 (10%) | 5 (6%) | 0.68 |
| Combined aortic and mitral valve | 15 (25%) | 99 (11%) | 93 (9%) | 70 (10%) | 0.001 A,B,C |
| Concomitant coronary artery bypass grafting | 14 (23%) | 241 (28%) | 320 (31%) | 20 (32%) | 0.16 |
| Concomitant atrial fibrillation ablation | 17 (28%) | 204 (24%) | 238 (23%) | 164 (24%) | 0.88 |
| Preoperative medications | |||||
| Angiotensin-converting enzyme inhibitor | 14 (23%) | 225 (26%) | 296 (29%) | 231 (33%) | 0.01 E |
| Clopidogrel | 4 (7%) | 16 (2%) | 20 (2%) | 14 (2%) | 0.1 |
| Aspirin | 16 (27%) | 225 (26%) | 229 (23%) | 183 (26%) | 0.19 |
| Warfarin | 9 (15%) | 51 (6%) | 54 (5%) | 37 (5%) | 0.02 A,B,C |
| Lipid lowering | 15 (25%) | 315 (36%) | 426 (42%) | 345 (50%) | <0.001 B,C,D,E |
| β Blocker | 33 (57%) | 411 (48%) | 541 (54%) | 425 (62%) | <0.001 E,F |
| Discharge medications ¶ | |||||
| Angiotensin-converting enzyme inhibitor | 12 (23%) | 191 (24%) | 260 (27%) | 193 (30%) | 0.05E |
| Clopidogrel | 1 (2%) | 54 (7%) | 69 (7%) | 57 (8%) | 0.18 |
| Aspirin | 50 (89%) | 774 (94%) | 945 (94%) | 622 (93%) | 0.33 |
| Warfarin | 37 (65%) | 487 (59%) | 528 (53%) | 350 (52%) | 0.01 D,E |
| Lipid lowering | 40 (71%) | 606 (73%) | 795 (79%) | 562 (84%) | <0.001 C,E,F |
| β Blocker | 44 (81%) | 651 (80%) | 841 (85%) | 558 (87%) | 0.004 D,E |
⁎ Defined as preoperative total cholesterol ≥200 mg/dl.
† Defined as preoperative blood pressure ≥140/90 mm Hg.
‡ Defined by preoperative patient questionnaire and confirmed using previous institutional medical records.
§ Included all first-degree relatives.
∥ Defined as angiographic evidence of obstructive lesion on formal cardiac catheterization.
| Variable | BMI (kg/m 2 ) | p Value | ||
|---|---|---|---|---|
| 30–34.9 | 35–39.9 | 40–60.5 | ||
| (n = 428) | (n = 173) | (n = 93) | ||
| Women | 157 (37%) | 79 (46%) | 52 (56%) | 0.001 A,B |
| Age (years) | 63.5 ± 13.4 | 62.0 ± 11.7 | 60.3 ± 12.2 | 0.06 |
| White | 333 (78%) | 132 (76%) | 63 (68%) | 0.12 |
| Current smoker | 19 (4%) | 6 (3%) | 0 (0%) | 0.11 |
| Diabetes mellitus | 88 (21%) | 59 (34%) | 49 (53%) | <0.001 A,B,C |
| Hyperlipidemia ⁎ | 260 (61%) | 108 (62%) | 62 (67%) | 0.56 |
| Renal failure | 26 (6%) | 9 (5%) | 8 (9%) | 0.54 |
| Dialysis | 7 (2%) | 4 (2%) | 6 (6%) | 0.024 B |
| Hypertension † | 308 (72%) | 122 (71%) | 78 (84%) | 0.041 B,C |
| Infective endocarditis | 21 (5%) | 8 (5%) | 6 (6%) | 0.79 |
| Chronic obstructive pulmonary disease | 53 (12%) | 30 (17%) | 23 (25%) | 0.008 B |
| Peripheral vascular disease | 33 (8%) | 13 (8%) | 8 (9%) | 0.95 |
| Cerebrovascular disease | 59 (14%) | 22 (13%) | 14 (15%) | 0.87 |
| Cerebrovascular accident | 49 (11%) | 19 (11%) | 11 (12%) | 0.98 |
| Previous myocardial infarction | 46 (11%) | 18 (10%) | 15 (16%) | 0.30 |
| Heart failure (by history) ‡ | 139 (32%) | 71 (41%) | 43 (46%) | 0.015 A,B |
| New York Heart Association class III-IV | 177 (42%) | 77 (45%) | 45 (48%) | 0.48 |
| Atrial fibrillation (by history) ‡ | 127 (30%) | 60 (35%) | 34 (37%) | 0.28 |
| Previous cardiovascular intervention | 140 (33%) | 53 (31%) | 35 (38%) | 0.51 |
| Angina pectoris | 63 (15%) | 30 (17%) | 15 (16%) | 0.72 |
| Family history of coronary artery disease § | 73 (17%) | 31 (18%) | 25 (27%) | 0.08 |
| Coronary artery disease ∥ | 183 (43%) | 62 (36%) | 41 (45%) | 0.19 |
| Repeat sternotomy | 68 (16%) | 33 (19%) | 17 (18%) | 0.60 |
| Elective surgery | 370 (86%) | 142 (82%) | 69 (74%) | 0.012 B |
| Urgent surgery | 53 (12%) | 29 (17%) | 22 (24%) | 0.017 B |
| Preoperative creatinine (mg/dl) | 0.022 C | |||
| Median | 1.0 | 1.0 | 1.1 | |
| Interquartile range | 0.9–1.3 | 0.8–1.2 | 0.9–1.4 | |
| Ejection fraction (%) | 0.85 | |||
| Median | 60 | 60 | 60 | |
| Interquartile range | 50–63 | 50–64 | 50–63 | |
| Perfusion time (min) | 0.09 | |||
| Median | 114 | 123 | 110 | |
| Interquartile range | 82–151 | 94–160 | 83–176 | |
| Cross-clamp time (min) | 0.30 | |||
| Median | 92 | 97 | 87 | |
| Interquartile range | 68–121 | 74–129 | 62–146 | |
| Underlying valve disease | ||||
| Aortic stenosis | 201 (49%) | 78 (47%) | 49 (57%) | 0.29 |
| Aortic regurgitation ≥2+ | 102 (25%) | 39 (23%) | 23 (26%) | 0.84 |
| Mitral stenosis | 33 (8%) | 20 (12%) | 12 (14%) | 0.18 |
| Mitral regurgitation ≥2+ | 168 (40%) | 64 (37%) | 36 (40%) | 0.79 |
| Tricuspid regurgitation ≥2+ | 69 (17%) | 25 (15%) | 11 (12%) | 0.53 |
| Surgery type | ||||
| Isolated valve | 140 (33%) | 51 (29%) | 33 (35%) | 0.58 |
| Aortic | 79 (18%) | 30 (17%) | 20 (22%) | 0.70 |
| Mitral | 39 (9%) | 14 (8%) | 8 (9%) | 0.92 |
| All aortic valve | 288 (67%) | 116 (67%) | 64 (69%) | 0.95 |
| Repair | 98 (34%) | 31 (27%) | 18 (28%) | 0.30 |
| Replacement | 190 (66%) | 85 (73%) | 46 (72%) | 0.30 |
| All mitral valve | 171 (40%) | 74 (43%) | 36 (39%) | 0.76 |
| Repair | 116 (68%) | 42 (57%) | 16 (44%) | 0.018 B |
| Replacement | 55 (32%) | 32 (43%) | 20 (56%) | 0.018 B |
| All tricuspid valve | 52 (12%) | 19 (11%) | 7 (8%) | 0.44 |
| Repair | 49 (94%) | 17 (89%) | 7 (100%) | 0.59 |
| Replacement | 3 (6%) | 2 (11%) | 0 (0%) | 0.59 |
| Combined aortic and mitral valve | 42 (10%) | 20 (12%) | 8 (9%) | 0.71 |
| Concomitant coronary artery bypass grafting | 136 (32%) | 55 (32%) | 29 (31%) | 0.99 |
| Concomitant atrial fibrillation ablation | 99 (23%) | 40 (23%) | 25 (27%) | 0.73 |
| Preoperative medications | ||||
| Angiotensin-converting enzyme inhibitor | 138 (32%) | 65 (38%) | 28 (30%) | 0.35 |
| Clopidogrel | 10 (2%) | 3 (2%) | 1 (1%) | 0.71 |
| Aspirin | 102 (24%) | 49 (29%) | 32 (34%) | 0.09 |
| Warfarin | 22 (5%) | 12 (7%) | 3 (3%) | 0.42 |
| Lipid lowering | 218 (51%) | 81 (47%) | 46 (50%) | 0.63 |
| β blocker | 258 (61%) | 105 (63%) | 62 (67%) | 0.50 |
| Discharge medications ¶ | ||||
| Angiotensin-converting enzyme inhibitor | 116 (29%) | 49 (30%) | 28 (34%) | 0.73 |
| Clopidogrel | 31 (7%) | 16 (10%) | 10 (11%) | 0.41 |
| Aspirin | 389 (94%) | 153 (92%) | 80 (91%) | 0.57 |
| Warfarin | 213 (51%) | 87 (52%) | 50 (57%) | 0.63 |
| Lipid lowering | 346 (83%) | 135 (80%) | 81 (92%) | 0.05 B,C |
| β Blocker | 352 (87%) | 133 (85%) | 73 (87%) | 0.71 |
⁎ Defined as preoperative total cholesterol level ≥200 mg/dl.
† Defined as preoperative blood pressure of ≥140/90 mm Hg.
‡ Defined according to preoperative patient questionnaire and confirmed using previous institutional medical records.
§ Included all first-degree relatives.
∥ Defined as angiographic evidence of obstructive lesion on formal cardiac catheterization.
The mean follow-up time was 35.8 ± 23.5 months (median 33.4, interquartile range 40.4). Both the underweight and the normal BMI groups had significantly greater (p <0.001) long-term mortality rates per 10 years of person-time (1.21 and 0.52, respectively) than the overweight (0.31) and obese (0.44) groups. During the entire follow-up duration, 338 total deaths occurred and were distributed across BMI categories as follows: underweight, 19; normal, 133; overweight, 99; obese class I, 58; obese class II, 15; and obese class III, 14. This same trend was observed for the incidence of in-hospital (p = 0.016) and operative (p = 0.035) mortality. Of the 91 total deaths (3.4%) that occurred during hospitalization, most were attributed to cardiac causes in each of the BMI categories, especially in patients who were underweight or obese ( Table 3 ). Underweight patients had a greater risk of long-term mortality than normal-weight patients (adjusted hazard ratio [HR] 1.69, 95% confidence interval [CI] 1.01 to 2.85, p = 0.048). Compared to normal BMI patients, overweight patients (adjusted HR 0.60, 95% CI 0.46 to 0.79, p <0.001) and obese patients (adjusted HR 0.67, 95% CI 0.50 to 0.91, p = 0.009) had a lower hazard for long-term all-cause mortality. Figure 1 displays the Kaplan-Meier survival curves stratified by BMI category (log-rank, p <0.001) for patients undergoing all valve surgeries. The weighted log-rank test indicated that differences in survival occurred early and persisted in the middle and late follow-up periods (p <0.001 for all analyses). Univariate Cox regression models revealed the following concordance index (C-index) values: BMI (categorical), 0.54 (95% CI 0.51 to 0.58); BMI (continuous), 0.54 (95% CI 0.51 to 0.58); age, 0.65 (95% CI 0.62 to 0.68); female gender, 0.54 (95% CI 0.51 to 0.57); and ejection fraction, 0.58 (95% CI 0.55 to 0.61). Figure 2 indicates that, when examined as a continuous variable, the BMI is associated in a nonlinear fashion with the natural logarithm of all-cause long-term mortality HR. In both unadjusted and covariate-adjusted models ( Figure 2 ), mortality risk was greatest for underweight patients, lower for the normal BMI group, lowest for the overweight patients, and slightly greater for the obese patient group. As revealed by the spline-based multivariate model, the following factors were significantly associated with increased all-cause long-term mortality, in addition to BMI: age (HR 1.03, p <0.0001), history of atrial fibrillation (HR 1.59, p = 0.008), hypercholesterolemia (HR 1.43, p = 0.015), chronic obstructive pulmonary disease (HR 1.86, p <0.0001), previous cardiac surgery (HR 1.69, p = 0.003), New York Heart Association class III-IV (HR 1.53, p = 0.001), and emergent surgical status (HR 1.75, p <0.0001).
| Cause of Death | BMI (kg/m 2 ) | All Valve (n = 2,640) | |||
|---|---|---|---|---|---|
| 11.5–18.4 | 18.5–24.9 | 25–29.9 | 30–60.5 | ||
| (n = 61) | (n = 865) | (n = 1,020) | (n = 694) | ||
| Cardiac | 3 (75%) | 20 (50%) | 13 (52%) | 18 (82%) | 54 (59%) |
| Heart failure | 0 | 0 | 1 (4%) | 0 | 1 (1%) |
| Infection | 0 | 5 (13%) | 3 (12%) | 1 (5%) | 9 (10%) |
| Intraoperative exsanguination | 0 | 1 (3%) | 0 | 0 | 1 (1%) |
| Multiorgan failure | 0 | 1 (3%) | 0 | 0 | 1 (1%) |
| Neurologic | 0 | 2 (5%) | 0 | 1 (5%) | 3 (3%) |
| Other | 0 | 0 | 0 | 1 (5%) | 1 (1%) |
| Pulmonary | 1 (25%) | 6 (15%) | 3 (12%) | 0 | 10 (11%) |
| Renal | 0 | 0 | 1 (4%) | 0 | 1 (1%) |
| Unknown | 0 | 4 (10%) | 3 (12%) | 0 | 7 (8%) |
| Vascular | 0 | 0 | 0 | 1 (5%) | 1 (1%) |
| Withdrawal of support | 0 | 1 (3%) | 1 (4%) | 0 | 2 (2%) |
| Total in-hospital deaths | 4 | 40 | 25 | 22 | 91 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
