Relationship Between Body Mass Index and Prognosis of Patients Presenting With Potential Acute Coronary Syndromes

Studies examining the relationship between obesity and acute coronary syndrome (ACS) have been limited to patients with confirmed diagnoses. The authors sought to determine the relationship between body mass index (BMI) and 30-day cardiovascular events in emergency department (ED) patients with potential ACS. This was a secondary analysis of a prospective cohort study of patients who presented to the ED with potential ACS. Patients were stratified according to their BMI: underweight (BMI < 18.49 kg/m2), normal weight (BMI = 18.5 to 24.99 kg/m2), overweight (BMI = 25 to 29.99 kg/m2), obese (BMI = 30 to 34.99 kg/m2), and very obese (BMI > 35 kg/m2). The primary outcome was acute myocardial infarction (AMI), death, or revascularization within 30 days of presentation. A logistic regression analysis was used to adjust for confounding variables and adjusted odds ratios (aOR) with 95% confidence intervals (CIs) are presented for cardiac events and readmission outcomes. Of the 3,946 patients included in this study, 73 (1.9%) were underweight, 911 (23%) were normal weight, 1,199 (30.4%) were overweight, 872 (22.1%) were obese, and 891 (22.6%) were very obese. Although increased levels of obesity were associated with a greater number of cardiac risk factors, there was no difference in 30-day cardiovascular events between those of normal weight and underweight (aOR = 1.1; 95% CI = 0.4 to 2.7), overweight (aOR = 1.0; 95% CI = 0.7 to 1.4), obese (aOR = 1.2; 95% CI = 0.8 to1.7), or very obese (aOR = 0.8; 95% CI = 0.5 to 1.3). Those who were underweight were more likely to be readmitted within 30 days (aOR = 1.9; 95% CI = 1.0 to 3.7), and those who were very obese were less likely to be readmitted within 30 days (aOR = 0.7; 95% CI = 0.5 to 0.9). Among patients who present to the ED with potential ACS, BMI is not associated with higher risk of cardiovascular outcomes at 30 days. Los estudios que examinan la relación entre la obesidad y el síndrome coronario agudo (SCA) se han limitado a pacientes con el diagnóstico confirmado. Los autores investigan la relación entre el índice de masa corporal (IMC) y los eventos cardiovasculares durante los 30 días siguientes en los pacientes con un potencial SCA que consultan en el servicio de urgencias (SU). Análisis secundario de un estudio de cohorte prospectivo de pacientes que acudieron al SU con un potencial SCA. Los pacientes fueron clasificados acorde a su IMC: peso bajo (IMC < 18,49 kg/m2), peso normal (IMC 18,5 a 24,99 kg/m2), sobrepeso (IMC 25 a 29,99 kg/m2), obeso (IMC 30 a 34.99 kg/m2) y muy obeso (BMI > 35 kg/m2). El resultado principal fue el infarto agudo de miocardio (IAM), la muerte o la revascularización en los primeros 30 días tras la visita. Se utilizó un análisis de regresión logística para ajustar las variables de confusión y se presentaron las razones de ventajas ajustadas (ORa) con los intervalos de confianza (IC) del 95% para los eventos cardiacos y los resultados de reingreso. De los 3.946 pacientes incluidos en este estudio, 73 (1,9%) fueron de bajo peso, 911 (23%) de peso normal, 1.199 (30,4%) con sobrepeso, 872 (22,1%) obesos y 891 (22.6%) muy obesos. Aunque los niveles elevados de obesidad se asociaron con un mayor número de factores de riesgo cardiovasculares, no hubo diferencias en los eventos cardiovasculares a los 30 días entre aquéllos con peso normal y bajo peso (ORa 1,1; IC 95% = 0,4 a 2,7), sobrepeso (ORa 1,0; IC 95% = 0,7 a 1,4), obesos (ORa 1,2; IC 95% = 0,8 a 1,7), o muy obesos (ORa 0,8; IC 95% = 0,5 a 1,3). Aquéllos que tenían bajo peso tuvieron mayor probabilidad de ser reingresados en los 30 días siguientes (ORa 1,9; IC 95% = 1,0 a 3,7), y aquéllos que fueron muy obesos tuvieron menor probabilidad (ORa 0,7; IC 95% = 0,5 a 0,9). En los pacientes que acuden al SU con un potencial SCA, el IMC no está asociado con un mayor riesgo de resultados cardiovasculares adversos a los 30 días. Every year, more than 6 million patients present to the emergency department (ED) with chest pain, and the majority are subsequently admitted with concern for acute coronary syndromes (ACS).1-3 As the annual expenditures for evaluation of ACS total $13 billion, continued efforts to determine methods that accurately risk-stratify patients may assist in differentiating between patients who need admission and those who can safely be discharged home.4-6 During the process of risk-stratifying patients with potential ACS, physicians often assess the presence of known cardiac risk factors.5, 7-9 Three of these risk factors, namely, hypertension, hypercholesterolemia, and diabetes mellitus, are all conditions that are strongly associated with obesity as well.10-12 The relationship between cardiovascular disease and obesity, a growing health concern that affects one out of every three Americans, has been the subject of extensive investigation.13 More specifically, studies have examined the role of body mass index (BMI), the height and weight quantification of body fat, in patients with confirmed ACS.14-25 Several of these studies have suggested that higher BMI was associated with more favorable short-term outcomes.15, 18-25 However, the effect of BMI on prognosis of patients with undifferentiated chest pain, the second leading cause for presentation to the ED, has not been previously described. This study aims to evaluate the relationship between BMI and 30-day cardiovascular outcomes in patients presenting to the ED with potential ACS. We hypothesized that elevated BMI would be associated with increased risk of 30-day cardiovascular events. We completed a secondary analysis of data from a prospective cohort study of patients presenting to the ED with potential ACS.26-30 The institutional review board approved the study. Subjects provided written informed consent. This study was conducted at the University of Pennsylvania, a tertiary-care teaching hospital located in an urban setting. The ED, which consists of 33 rooms and an eight-room lower-acuity unit, has an annual census of over 60,000 patients. Trained research assistants (RAs) staffed the ED 7 days a week from the hours of 7:00 a.m. through 12:00 a.m., and were responsible for the screening and enrollment processes of all patients eligible for the study.31 Study participants comprised a convenience sample of ED patients enrolled over a 5-year period. Patients included were those over 18 years of age who presented to the ED with symptoms that may indicate potential ACS and who received electrocardiograms (ECG). Patients were excluded if symptoms at presentation were the result of clearly noncardiac etiologies, such as trauma. Patients who were unable to provide informed consent, such as those who were not English-speaking or were cognitively impaired, were also excluded from participation. Last, enrollment in this study had no effect on the clinical management of each patient. All decisions related to care were at the discretion of the treating physician. Data collection was performed using the Standardized Reporting Guidelines for Emergency Department Chest Pain Studies.32 The treating emergency physician recorded all of the following information for each participant: pain character and location, associated symptoms, cardiac risk factors, past medical history, prior cardiac interventions, and cardiac testing, Thrombosis In Myocardial Infarction (TIMI) risk score values, cardiac findings from physical examination, and interpretation of the initial ECG.32, 33 RAs obtained demographic and contact information directly from each participant and recorded data noted by triage staff that included initial vital signs, medication use, and triage class. Throughout the duration of a participant's ED stay, the RA also documented biomarker values, treatments, imaging results, and final disposition. Research assistants tracked the hospital courses of all study participants who were admitted to the hospital. Any medical intervention, diagnostic test result, or cardiovascular event that occurred in house was recorded. The final diagnoses upon discharge were also recorded. After 30 days following the index visit, an RA contacted each patient or a patient proxy via telephone and reviewed medical records as needed to document outcomes. Rehospitalization and occurrences of cardiovascular events were recorded, as well as results from cardiac testing or procedures that were completed subsequent to the index visit. Patients were stratified into five BMI groups similar to the World Health Organization guidelines for BMI classification.34 Those with BMI values under 18.49 kg/m2 were classified as underweight, those with values between 18.5 and 24.99 kg/m2 were described as being within normal weight range, those with values between 25 and 29.99 kg/m2 were defined as overweight, those with values between 30.00 and 34.99 kg/m2 were described as obese, and those with values greater than 35 kg/m2 were classified as being very obese. The primary measure of outcome was a composite of acute myocardial infarction (AMI), revascularization, and death within 30 days of presentation. All diagnoses of AMI were defined by and had cardiac biomarker levels consistent with criteria established by the European Society of Cardiology and the American College of Cardiology.35 Revascularization procedures included percutaneous coronary intervention or coronary artery bypass grafting. Outcomes were recorded through medical record review and follow-up telephone conversation with each participant. In cases where research staff were unable to contact a patient, communication was attempted with a patient proxy, family member, or treating physician. RAs consulted the Social Security Death Index (SSDI) as an additional resource to assess patient survival. SSDI search was performed at least 2 years after presentation to increase the level of accuracy of this search. Outcomes involving cardiovascular events were confirmed through review of patient medical records. The secondary outcome was rate of readmission to the hospital within 30 days of the index visit. This outcome was determined using the same method of follow-up as for the primary measure, and readmissions were confirmed by review of medical records. Study data were entered into a Microsoft Access database (2003 version, Microsoft Corp., Redmond, WA). Data for this study were extracted from the database and were processed with STATA software (version 12, StataCorp, College Station, TX) for statistical analysis. Data comparing the different strata of BMI with respect to past history and risk factors were not formally compared, as there would be expected to be differences. These data are presented descriptively. We performed a direct logistic regression to determine association between BMI strata and 30-day cardiovascular events controlling for the following prespecified covariates: age, sex, diabetes, hypertension, family history of MI, tobacco use, hypercholesterolemia, known coronary artery disease (CAD), prior AMI, presence of ST-segment deviation, left ventricular hypertrophy, T-wave inversions, hyperacute T waves, pathologic Q waves, and bundle branch blocks. These variables were selected a priori based on review of prior literature for risk factors that are associated with 30-day outcome and obesity, to reduce potential confounding.12, 19, 23, 26 As a rule of thumb, we used the recommendations by Peduzzi et al.36 and allowed for 10 events per variable. We had 294 thirty-day events and could include our prespecified covariates in our model without risking the validity of our model. Adjusted odds ratios (aOR) are presented with 95% confidence interval (CI), using the normal weight range as the reference group. Influential outliers were assessed using a combination of visual inspection, Pearson residuals, and Pregibon delta beta and when removed from the final models were not found to change the inference. The greatest change in point estimates was observed in the underweight category, where five observations were identified as influential (total underweight n = 73), all of which recorded at least one outcome (total outcomes in underweight category n = 6). Multicollinearity was assessed using variance inflation factor values and no issues were detected. The assumption of linearity was evaluated in the variable for patient age in our final model. Transformation of age resulted in no significant change in inference. Goodness of fit of the overall model was assessed using the Hosmer-Lemeshow goodness-of-fit statistic. A total of 4,492 patients were enrolled in the primary study. After exclusion of 110 patients who were missing data necessary for the BMI calculation, and 267 who had recent cocaine use, there were 3,946 patients included in this analysis. Among these patients, 73 were underweight (1.9%), 911 were normal weight (23%), 1199 were overweight (30.4%), 872 (22.1%) were obese, and 891 (22.6%) were very obese. The average mean (±SD) age was 53.8 (±14.5) years, 57% were men, and 65% were African American. Additional distribution of baseline characteristics such as cardiac risk factors and past cardiac history are reported in Table 1. Of the 3,946 patients in this analysis, 2,770 (70.2%) were admitted to the hospital. During hospitalization, 248 patients (6.3%) experienced cardiovascular events, 141 (3.6%) of whom sustained AMI, 154 (3.9%) of whom underwent revascularization, and 27 (0.7%) of whom died (Table 2). Overweight patients exhibited the highest rate of AMI and, similar to obese patients, were more likely to undergo revascularization when compared to other BMI groups. Underweight patients exhibited the highest mortality (Table 2). During the 30 days following hospitalization, of the 3,946 patients in this analysis, 11 patients (0.3%) sustained AMI, 23 (0.6%) underwent revascularization, and 30 (0.7%) died (Table 2). Occurrence rates of AMI and revascularization were comparable among BMI groups. Underweight patients and normal weight patients exhibited the highest mortality rates (Table 2). Overall, the occurrence rates for the three types of cardiac events during hospitalization and at 30 days were not significantly different across BMI groups. In unadjusted analyses, there was no association of BMI with cardiac outcomes. After adjusting for age; sex; diabetes; hypertension; family history; tobacco use; hypercholesterolemia; known CAD; prior MI; and presence of ST-segment deviation, left ventricular hypertrophy, T wave inversions, hyperacute T waves, pathologic Q waves, and bundle branch blocks, there was no association between BMI and 30-day cardiovascular events (Table 3). The Hosmer-Lemeshow goodness of fit of the model had a p-value of 0.73. We were unable to obtain reliable data regarding rehospitalization in 324 subjects (8%), leaving 3,622 patients (92%) for the secondary outcome assessment. Among these patients, 426 patients (11.8%) were readmitted within 30 days subsequent to their index visits. Readmission rates were similar among normal weight, overweight, and obese patients. In contrast, underweight and very obese patients exhibited the highest and lowest readmission rates, respectively (Table 4). Logistic regression modeling was used to evaluate readmission outcomes and BMI. Adjusted readmission rates were highest in underweight patients (aOR = 1.9; 95% CI = 1.0 to 3.7), whereas very obese patients (aOR = 0.7; 95% CI = 0.5 to 0.9) had lower risk of readmission (Table 5). The Hosmer-Lemeshow goodness of fit had a p-value of 0.38. To our knowledge, our study is the first to report the relationship between BMI and outcomes in patients with potential ACS. Prior reports have been limited to high-risk patients with confirmed diagnoses of ACS. Several of these studies support the notion of an “obesity paradox,” a phenomenon where elevated BMI is associated with better short-term outcomes when compared to normal weight patients.15, 18, 23 Other findings suggest a U-shaped relationship, whereby overweight and obese patients exhibit the most favorable outcomes.16, 19-21 For example, Diercks et al.19 found that in 80,845 patients with non-ST-segment elevation ACS, obese and overweight patients had lower risk for AMI and death during the hospital stay when compared to normal weight patients. Conversely, underweight and very obese patients had the worst outcomes.19 In an investigation of 50,149 patients with ST-segment elevation myocardial infarction, Das and colleagues23 demonstrated that extreme obesity was associated with higher patient mortality. Other studies have presented divergent findings, and some authors posit that any protective effects of obesity result from confounding; thus the relationship between BMI and outcomes remains unclear.16, 17, 23, 24 We believe that our analysis might provide a meaningful contribution that extends existing knowledge. Our study encompassed a wider range of patients, namely, those with undifferentiated chest pain, the second leading cause of presentation to the ED.36 Moreover, the majority of patients in this analysis represented a low-risk cohort, which is relevant given that low- to intermediate-risk patients account for 50% to 70% of all potential ACS presentations.37 Nevertheless, we acknowledge that higher BMI groups may be fundamentally different in ways not included in our analysis. We did not adjust for pharmacologic treatment, clinical interventions, or comorbid conditions, which may explain certain outcomes. Although we did adjust for several confounding factors, we were unable to clearly define what may have affected readmission rates. A common finding in confirmed ACS studies suggests that obese patients were usually younger and treated more aggressively, possibly contributing to improved outcomes.16, 17, 23 Other authors have reasoned that better outcomes in obese patients may be due to larger coronary artery diameter, higher levels of inflammatory factors, and cautious patient selection for interventions or more aggressive management.19 In general, underweight patients have consistently demonstrated worse clinical outcomes, a this has to be reports suggest that underweight patients may be at higher risk due to comorbid and greater risk for pharmacologic have that used to ACS may underweight patients and may be limited with respect to or may cause greater in these patients in is a risk factor for cardiovascular disease and is strongly associated with comorbid conditions such as hypertension, diabetes mellitus, and hypercholesterolemia, all of which are factors for cardiac events. these might to or management of ACS is not et sought to the number of cardiac risk factors might assist a physician in a patient with potential ACS. In a sample of patients, found that cardiac risk factor clinical in the of ACS in patients than years of In this and with reported findings, we would that obesity to higher risk and cardiovascular events, this may not assist in short-term risk in patients presenting to the ED with potential ACS. we believe that obesity may not assist in overall for ED of patients with potential Our analysis on previously and confounding variables may have to the observed relationship between BMI and outcomes. Although we adjusted for variables such as age, sex, cardiac risk factors, and medical history, we were unable to for related to risk or of We also could not adjust for level and of of of body or all of which have been with cardiovascular The of BMI as the to obesity within our analysis. height and weight are thus there may be or Several studies have suggested that and which were not in this study, are to BMI in cardiovascular risk and outcomes following The values used to BMI were obtained during the index visits. the index in such as or also an account of in BMI over However, et in that weight is more likely to the short-term follow-up in our study, in BMI for several patients are Patients included in this analysis represented a convenience sample that may have been subject to selection Although the primary study attempted to this with more subject enrollment was also upon the of research the study was African with of and other This is a study of a African and thus our results may not be to other or Other may the outcome in this study. outcomes were obtained through communication with a patient or proxy, for In an to this outcome were confirmed with medical records when However, was that study subjects could have sustained cardiovascular events or readmission at without our knowledge. The potential effect of obesity relevant in the clinical setting. We determined that body mass index was not associated with 30-day cardiovascular outcomes in patients who presented to the ED with potential acute coronary