In patients with hypertension, heart failure, or coronary artery disease (CAD), obese patients have been shown to have a lower cardiac event rate compared with normal weight counterparts. This phenomenon has been termed the “obesity paradox.” We sought to determine whether the obesity paradox exists in a cohort of patients referred for stress echocardiography. We evaluated 4,103 patients with suspected CAD (58 ± 13 years; 42% men) undergoing stress echocardiography (52% exercise and 47% dobutamine). Patients were divided into 3 groups on the basis of body mass index (BMI): 18.5 to 24.9, 25 to 29.9, and >30 kg/m 2 . During the follow-up of 8.2 ± 3.6 years, there were 683 deaths (17%). Myocardial ischemia was present in 21% of the population. Myocardial ischemia was more prevalent in patients with a BMI of 18.5 to 24.9 kg/m 2 (26%) than those with a BMI of 25 to 29.9 kg/m 2 (21%) and >30 kg/m 2 (18%). Patients with a BMI of >30 kg/m 2 had the lowest death rate (1.2%/year) compared with those with a BMI of 25 to 29.9 kg/m 2 (1.75%/year) and 18.5 to 24.9 kg/m 2 (2.9%/year; p <0.001). After adjusting for significant clinical variables including exercise capacity, patients with higher BMI (>30 kg/m 2 and 25 to 29.9 kg/m 2 ) had less risk of mortality compared with those with a BMI of 18.5 to 24.9 kg/m 2 (hazard ratio 0.58, 95% confidence interval 0.47 to 0.72, p <0.0001 and hazard ratio 0.69, 95% confidence interval 0.57 to 0.82, p <0.0001, respectively). In conclusion, higher survival rate in patients with higher BMI as previously described in patients with hypertension, heart failure, and CAD extends to patients with suspected CAD referred for stress echocardiography, independent of exercise capacity.
The detrimental effect of obesity on morbidity and mortality has been demonstrated in numerous epidemiological studies. It has been estimated that approximately, 1 in 10 deaths in the United States can be attributed to obesity in the United States population. Studies have also demonstrated a significantly greater prevalence of coronary artery disease (CAD) risk factors including dyslipidemia, hypertension, diabetes, and heart failure in obese patients. With the increased prevalence of CAD risk factors seen in obese patients, it would logically be concluded that obesity should increase mortality in patients with known CAD. In 2002, however, Gruberg et al showed that patients with lower body mass index (BMI) had greater risk of cardiac death after a percutaneous coronary intervention. Similar counterintuitive results have been demonstrated in other cardiovascular pathologies such as acute myocardial infarction, heart failure, and hypertension as well as patients with peripheral artery disease and after cardiac surgery.
Stress echocardiography is clinically indicated for risk stratification, prognosis, and management of patients with suspected or known CAD. We evaluated the impact of BMI on prognostic outcomes in patients referred for stress echocardiography.
Methods
We identified 4,103 consecutive patients referred for exercise or pharmacologic stress echocardiography during March 1997 to June 2010 at St. Luke’s-Roosevelt Hospital Medical Center, Columbia University Hospital of Physicians and Surgeons, New York, New York. Patients were excluded if they had previous myocardial infarction, previous percutaneous intervention or coronary artery bypass graft; a history of heart failure; poor acoustic windows (<13 of 16 segments visualized by echocardiography) despite image-enhancing agents; patients with a BMI of <18.4 kg/m 2 . BMI was calculated as weight in kilograms divided by height in meters squared. Patients were divided into 3 groups on the basis of BMI (kg/m 2 ): 18.5 to 24.9, 25 to 29.9, and >30 kg/m 2 . Patients who reached ≥6 METs on exercise were considered to have good exercise capacity (GEC) and those who attained <6 METs were considered to have poor exercise capacity (PEC). We used this approach to maintain consistency in our study methods and because a widely accepted clinical categorization of fitness does not exist. Patients were divided into 9 groups based on BMI, exercise capacity, and stress mode: BMI of 18.5 to 24.9 kg/m 2 and GEC; BMI of 25 to 29.9 kg/m 2 and GEC; BMI of >30 kg/m 2 and GEC; BMI of 18.5 to 24.9 kg/m 2 and PEC; BMI of 25 to 29.9 kg/m 2 and PEC; BMI of >30 kg/m 2 and PEC; BMI of 18.5 to 24.9 kg/m 2 and pharmacologic stress; BMI of 25 to 29.9 kg/m 2 and pharmacologic stress; and BMI of >30 kg/m 2 and pharmacologic stress. This study was approved by St. Luke’s-Roosevelt Hospital Institutional Review Board.
Exercise was the preferred stress method in patients who were able to exercise to an adequate workload, which was defined as at least 85% of age-adjusted maximal predicted heart rate. Maximal exercise treadmill testing was performed using a standard Bruce protocol. Dobutamine was administered intravenously beginning at a dose of 5 to 10 μg/kg/min and increased by 5 to 10 μg/kg every 3 minutes up to a maximum of 50 μg/kg/min, or until a study end point was achieved. The end points for termination of the dobutamine infusion included development of new segmental wall motion abnormalities, attainment of >85% of age-predicted maximum heart rate, or the development of significant adverse effects related to the dobutamine infusion. Atropine was administered intravenously in 0.25 to 0.5 mg increments up to a maximum dose of 2.0 mg if a study end point was not achieved.
Beta blockers were held on the morning of the test, as is the protocol in our laboratories. Seven standard echocardiographic views were obtained with each acquisition. The acquisition sequences were as follows: for each patient, the apical 4-chamber was acquired first, followed by 3-chamber, 2-chamber, and then parasternal long- and short-axis views (typically <30 to 40 seconds). This was then followed by the subcostal views (typically within the next 5 to 10 seconds). All images for exercise stress echocardiography were acquired within 30 to 60 seconds, and the entire sequence was repeated again to acquire a second run of images. During dobutamine stress echocardiography, images were acquired at baseline, with each increment of dobutamine infusion, and during the recovery phase. Cardiac rhythm was monitored throughout the stress echocardiography protocol, and 12-lead electrocardiograms and blood pressure measurements were obtained at baseline, at each level of stress, and during the recovery phase. Transthoracic echocardiographic images were obtained using commercially available ultrasound equipment (Acuson Sequoia, Mountain View, California and General Electric Vivid 7 and 9, Milwaukee,Wisconsin).
The left ventricle was divided into 16 segments as recommended by the American Society of Echocardiography, and a score was assigned to each segment at baseline, with each stage of stress (dobutamine only), and during recovery. Each segment was scored as follows: 1 = normal; 2 = mild to moderate hypokinesis (reduced wall thickening and excursion); 3 = severe hypokinesis (markedly reduced wall thickening and excursion); 4 = akinesis (no wall thickening and excursion); 5 = dyskinesis (paradoxical wall motion away from the center of the left ventricle during systole). All echocardiograms were interpreted by an experienced echocardiographer.
A normal response to stress was defined as normal wall motion at rest with increase in wall thickening and excursion during stress. An abnormal response to stress was defined as either deterioration in left ventricular wall segment thickening and excursion during stress (increase in wall motion score of ≥1 grade) or biphasic response with dobutamine stress. Ejection fraction at rest was visually estimated. Each segment was scored as either ischemic, scar, or viable with inotropic contractile reserve. For patients who underwent exercise, wall motion abnormality at rest without any improvement at peak stress was considered ischemic segment. Severity of ischemia was calculated by worsening of the peak wall motion score from normal. For example, a patient with ischemia in 2 segments, if the peak wall motion score for one segment is 3 (severe hypokinesis) and the peak wall motion score of the other segment is 4 (akinesis), and because a score of 1 indicates normal, the ischemic severity score is (3−1) + (4−1) = 5. A ischemic severity score of 0 indicates no ischemia. The maximal ischemic severity score for a patient could be 4 × 16 = 64.
Patients were followed up for a mean of 8.2 ± 3.6 years, with a maximum follow-up of 12 years. All-cause mortality was assessed using Social Security Death Index.
All analyses were carried out using a standard statistical package (SPSS, version 16.0; IBM, Armonk, New York). Continuous variables are expressed as means ± SD. Categorical variables are expressed as number and percentage. Univariate analyses of categorical variables were performed with chi-square test. One-way analysis of variance with a post hoc Boneferroni test was used to compare means of continuous variables among multiple groups. Annual event rates were calculated using the life table survival analyses. Univariate analyses associated with all-cause mortality were calculated using chi-square and relative hazards for all-cause mortality considering the time of death using Cox proportional analyses. Adjusted multivariate analysis was performed using the Cox proportional hazard including age, gender, CAD risk factors, BMI, ejection fraction, heart rate at rest, results of stress echocardiography, and modes of stress and fitness. An interaction of BMI and stress mode and/or exercise capacity was included in the model to determine any interaction. Cox proportional hazard analyses were applied to estimate adjusted hazard ratios and 95% confidence intervals for potential predictors of survival. Multivariate adjusted survival curves according to BMI were plotted by Cox proportional hazard analyses. Proportional hazard assumptions were tested with Schoenfeld residuals for continuous variables, log minus log for categorical variables, and time-dependant covariate for continuous and categorical variables. p Values were considered significant at <0.05.
Results
Patients’ demographics are characterized in Table 1 . Of the 4,103 patients studied, 1,722 (42%) were men, and the mean age was 58 ± 18 years. When classified by BMI, 1,043 (25%) had a BMI of 18.5 to 24.9 kg/m 2 , 1,430 (35%) had 25 to 29.9 kg/m 2 , and 1,630 (40%) had >30 kg/m 2 . Patients with BMIs of 25 to 29.9 kg/m 2 and >30 kg/m 2 were significantly younger than patients with a BMI of 18.5 to 24.9 kg/m 2 . More women than men were present in all BMI subgroups. About 75% of our study population comprised African-Americans and Hispanics. Patients with a BMI of >30 kg/m 2 had a higher percentage of metabolic risk factors such as hypertension, diabetes mellitus, hyperlipidemia, and family history of early CAD compared with patients with a BMI of 18.5 to 24.9 kg/m 2 who had a higher history of smoking. Patients with high BMI (>30 kg/m 2 ) were prescribed more angiotensin-converting enzyme inhibitors and angiotensin receptor blockers, diuretics, and lipid-lowering medications. There were no differences in antiplatelets (including aspirin and clopidogrel), β blockers, nitrates, and calcium channel inhibitors among all 3 groups.
Variable | BMI (kg/m 2 ) | p | |||
---|---|---|---|---|---|
18.5–24.9 (n = 1,043) | 25–29.9 (n = 1,430) | >30 (n = 1,630) | Total (n = 4,103) | ||
Age (yrs) | 62 ± 15 | 59 ± 23 | 55 ± 13 | 58 ± 18 | <0.0001 |
Men | 494 (47) | 692 (49) | 536 (33) | 1,772 (42.3) | <0.0001 |
Race | <0.001 | ||||
Caucasian | 254 (24) | 236 (17) | 212 (13) | 702 (17) | |
African-American | 403 (39) | 558 (39) | 734 (45) | 1,695 (41) | |
Hispanic | 282 (27) | 544 (38) | 587 (36) | 1,413 (34) | |
Asian and others | 104 (10) | 92 (6) | 97 (6) | 293 (7) | |
Hypertension ∗ | 564 (55) | 892 (63) | 1,143 (71) | 2,599 (64) | <0.0001 |
Diabetes mellitus | 203 (20) | 378 (27) | 564 (35) | 1,145 (28) | <0.0001 |
Family history of premature CAD | 221 (22) | 316 (22) | 418 (26) | 955 (23.5) | 0.018 |
Dyslipidemia † | 326 (32) | 569 (40) | 699 (43) | 1,594 (39) | <0.0001 |
Smoking tobacco (active) | 267 (26) | 323 (23) | 331 (21) | 921 (22) | 0.018 |
Symptoms ‡ | <0.001 | ||||
No chest pain or SOB | 363 (35) | 407 (29) | 471 (29) | 1,241 (30) | |
Nonanginal chest pain | 45 (4) | 62 (4) | 55 (3) | 163 (4) | |
Atypical angina | 454 (44) | 686 (48) | 704 (43) | 1,844 (45) | |
Typical angina | 98 (9) | 141 (10) | 179 (11) | 418 (10) | |
SOB | 83 (8) | 133 (9) | 221 (14) | 437 (11) | |
SOB | 149 (14) | 242 (17) | 307 (19) | 698 (17) | 0.009 |
Antiplatelet agents | 320 (31) | 461 (33) | 506 (32) | 1,287 (32) | 0.69 |
Nitrates | 50 (5) | 76 (6) | 62 (4) | 188 (5) | 0.134 |
β Blocker | 249 (25) | 345 (25) | 422 (27) | 1,016 (26) | 0.39 |
Calcium channel blocker | 169 (17) | 210 (15) | 282 (18) | 661 (17) | 0.13 |
ACE-I or ARB | 211 (21) | 340 (24) | 438 (28) | 989 (25) | 0.001 |
Diuretic | 119 (12) | 210 (15) | 340 (21) | 669 (17) | <0.0001 |
Lipid lowering | 168 (17) | 306 (22) | 338 (21) | 812 (20) | 0.003 |
All-cause mortality | 275 (26) | 226 (16) | 182 (11) | 683 (17) | <0.001 |
Stress mode and fitness | <0.0001 | ||||
≥6 METs | 518 (50) | 770 (54) | 668 (41) | 1,956 (48) | |
<6 METs | 45 (4) | 74 (5) | 98 (6) | 217 (5) | |
Pharmacologic | 480 (46) | 586 (41) | 864 (53) | 1,930 (47) | |
Heart rate at rest (beats/min) | 72 ± 13 | 72 ± 13 | 74 ± 14 | 73 ± 14 | <0.0001 |
Systolic blood pressure at rest (mm Hg) | 130 ± 22 | 133 ± 21 | 134 ± 21 | 133 ± 21 | <0.0001 |
Diastolic blood pressure at rest (mm Hg) | 77 ± 11 | 78 ± 11 | 78 ± 11 | 78 ± 12 | <0.0001 |
Left ventricular ejection fraction (%) | 56 ± 9 | 57 ± 8 | 58 ± 7 | 57 ± 8 | <0.0001 |
Ejection fraction >55% | 923 (89) | 1,288 (90) | 1,529 (94) | 3,740 (91) | <0.0001 |
Dilated left atrium (≥4 cm) | 421 (40) | 599 (42) | 812 (50) | 1,832 (45) | <0.0001 |
Left atrium size index | 2.2 ± 0.4 | 2.1 ± 0.3 | 1.9 ± 0.3 | 2.0 ± 0.4 | <0.0001 |
Rest wall motion index | 1.12 ± 0.4 | 1.09 ± 0.4 | 1.06 ± 0.3 | 1.08 ± 0.4 | <0.001 |
Stress wall motion index | 1.14 ± 0.4 | 1.09 ± 0.3 | 1.07 ± 0.3 | 1.1 ± 0.3 | <0.001 |
Presence of ischemia | 278 (27) | 305 (21) | 288 (18) | 871 (21) | <0.001 |
Total number of ischemic segments | 1.08 ± 2.3 | 0.8 ± 1.9 | 0.6 ± 1.8 | 0.8 ± 2.0 | <0.001 |
Total number of scar segments | 0.25 ± 1.3 | 0.19 ± 1.2 | 0.12 ± 0.9 | 0.2 ± 1.2 | 0.02 |
Ischemic burden | 1.5 ± 3.5 | 1.0 ± 2.7 | 0.9 ± 3.0 | 1.05 ± 3.1 | <0.001 |
∗ History of hypertension and/or use of antihypertensive medications.
‡ Patients who had both chest pain and shortness of breath were considered in the chest pain group.
Patients with high BMI (>30 kg/m 2 ) were more likely to undergo a pharmacologic stress test. The mean ejection fraction at rest for the cohort was 57 ± 8%. Table 2 summarizes exercise characteristics according to BMI. Maximum predicted heart rate achieved during exercise was adequate in all BMI groups. Exercise duration and METs achieved were lesser for patients with high BMI (>30 kg/m 2 ). Duke treadmill score was significantly higher in patients with a BMI of 18.5 to 24.9 kg/m 2 (7.6 ± 4.3) and 25 to 29.9 kg/m 2 (7.7 ± 4.0) compared with those with a BMI of >30 kg/m 2 (6.4 ± 3.5). Ischemia was present in 21% of the population. Patients with high BMI (25 to 29.9 and >30 kg/m 2 ) had fewer abnormal stress test results compared with those with a BMI of 18.5 to 24.9 kg/m 2 (21% vs 18% vs 26%, respectively; p <0.001; Table 1 ). Similarly, wall motion index at rest, stress wall motion index, extent and severity of ischemia were higher in patients with normal BMI (18.5 to 24.9 kg/m 2 ) compared with those with higher BMI ( Table 1 ). Dilated left atrium was seen in 45% of our cohort. Patients with higher BMI were more likely to have dilated left atrium; however, when indexed to body surface area, patients with a BMI of 18.5 to 24.9 kg/m 2 had higher left atrial size index.
Variable | BMI (kg/m 2 ) | p | |||
---|---|---|---|---|---|
18.5–24.9 (n = 563) | 25–29.9 (n = 844) | >30 (n = 766) | Total (n = 2,173) | ||
Peak heart rate (beats/min) | 157 ± 21 | 160 ± 18 | 159 ± 18 | 159 ± 19 | 0.04 |
Percent maximum predicted heart rate achieved | 95 ± 11 | 96 ± 10 | 94 ± 9 | 95 ± 10 | 0.007 |
Peak systolic blood pressure (mm Hg) | 155 ± 23 | 161 ± 23 | 163 ± 25 | 160 ± 24 | <0.0001 |
Peak diastolic blood pressure (mm Hg) | 80 ± 11 | 83 ± 12 | 84 ± 12 | 83 ± 12 | <0.0001 |
METs | 10.2 ± 3.4 | 9.9 ± 3.0 | 8.7 ± 2.6 | 9.5 ± 3.0 | <0.0001 |
Exercise duration (minutes) | 8.7 ± 3.1 | 8.5 ± 3.1 | 7.3 ± 2.5 | 8.1 ± 3.0 | <0.0001 |
Exercise chest pain | 14 (3) | 10 (1) | 11 (1) | 36 (2) | 0.3 |
Exercise ST response | 0.12 | ||||
Normal | 420 (75) | 624 (74) | 582 (76) | 1,626 (74) | |
Nondiagnostic | 58 (10) | 110 (13) | 92 (12) | 260 (12) | |
Equivocal (0.5–0.9 mm) | 23 (4) | 44 (5) | 24 (3) | 91 (4) | |
Ischemic >1 mm | 62 (11) | 66 (8) | 68 (9) | 196 (9) | |
Duke treadmill score | 7.6 ± 4.3 | 7.7 ± 4.0 | 6.4 ± 3.5 | 7.2 ± 4.0 | <0.0001 |