Presented at the Research Forum, American College of Emergency Physicians, San Francisco CA, October 2011.
Original Research Contribution
Lack of Sex Disparity in Cardiovascular Testing After Coronary Computerized Tomographic Angiography
Version of Record online: 9 FEB 2012
© 2012 by the Society for Academic Emergency Medicine
Academic Emergency Medicine
Volume 19, Issue 2, pages 147–152, February 2012
How to Cite
Ginty, C. T., Chang, A. M., Matsuura, A. C., Decker, C., Le, J., Green, M., Litt, H. I. and Hollander, J. E. (2012), Lack of Sex Disparity in Cardiovascular Testing After Coronary Computerized Tomographic Angiography. Academic Emergency Medicine, 19: 147–152. doi: 10.1111/j.1553-2712.2011.01286.x
Disclosures: Dr. Hollander—Abbott, Allere, Biosite, Brahms, BMS, Emergency Care Education and Consulting LLC, FA Davis, Nanosphere, Sanofi-Aventis, Siemens, Up-to-Date, the Duke Clinical Research Institute, the Pennsylvania Department of Health, the NIH, and the Annals of Emergency Medicine. Dr. Litt—Medrad/Bayer AG, Siemens Medical Solutions. Drs. Chang, Ginty, Matsuura, and Mr. Green disclosed no conflicts of interest. Mr. Decker and Mr. Le did not respond to the request for disclosure.
Supervising Editor: Brian Hiestand, MD.
A related commentary appears on p. 197.
- Issue online: 9 FEB 2012
- Version of Record online: 9 FEB 2012
- Received June 13, 2011; revisions received August 16 and August 23, 2011; accepted August 23, 2011.
ACADEMIC EMERGENCY MEDICINE 2012; 19:147–152 © 2012 by the Society for Academic Emergency Medicine
Objectives: The authors assessed whether there was a sex disparity in testing of patients after coronary computerized tomographic angiography (CTA) was performed for emergency department (ED) patients with potential acute coronary syndromes (ACS). In theory, once coronary anatomy has been determined, any disparity in subsequent workup should not be the result of differences in presentation.
Methods: This was a prospective cohort study of ED patients who presented with potential ACS and received coronary CTAs at a university hospital. Demographics, history, cardiac risk factors, follow-up testing, and procedures were recorded. Follow-up at 30 days was obtained by structured record review and telephone contact. Patients were stratified by sex and coronary CTA results (max stenosis: none, 1% to 24%, 25% to 49%, 50% to 69%, and ≥70%). Main outcome was the relative risk (RR) of a male receiving a stress test or catheterization within 30 days, stratified by categories of percent maximal stenosis.
Results: A total of 1,144 patients received coronary CTAs (mean ± SD age = 47.8 ± 8.7 years), 55% were female, and 64% were black or African American. Overall, 161 patients received follow-up testing within 30 days, 113 during their index visit. Men were more likely to receive further testing (RR = 1.51; 95% confidence interval [CI] = 1.14 to 1.99) compared to women. However, when stratified by percentage of stenosis, men were not more likely to receive further testing within 30 days after coronary CTA compared to women (RR = 1.14; 95% CI = 0.68 to 1.91). In multivariable modeling for risk of further testing, stenosis remained significant (adjusted relative risk [aRR] = 1.51; 95% CI = 1.19 to 1.91), while male sex, age, race, and Thrombolysis in Myocardial Infarction (TIMI) risk score were not.
Conclusions: Male patients with potential ACS who receive a coronary CTA as a part of their ED evaluation were no more likely than female patients to receive further testing within 30 days.
In the United States, nearly 2,300 individuals die of cardiovascular disease each day, the equivalent of one death every 38 seconds.1 According to 2006 data, 47% of deaths due to coronary heart disease occurred in women, making it the leading cause of death among women in the United States.1,2
Despite these figures, a sex disparity exists in cardiovascular care in this country. Several studies have shown that women receive fewer invasive procedures, such as cardiac catheterization, percutaneous coronary interventions (PCIs), and coronary artery bypass graft (CABG) compared to men, even when they present similarly.2,3 In patients presenting to the emergency department (ED) with potential acute coronary syndrome (ACS), men were nearly twice as likely as women to receive a diagnostic cardiac catheterization and more likely to receive noninvasive testing even after adjustment for factors such as age, Thrombolysis in Myocardial Infarction (TIMI) score,4,5 and cardiac risk factors.6 Several potential explanations for this discrepancy have been proposed, including sex-specific patient preferences,7 physician bias, or an overtreatment of men at low risk of disease and death.7–9 Unlike many of the most commonly used noninvasive diagnostic tests, coronary computerized tomographic angiography (CTA) has been shown to have nearly equal accuracy in men and women.2,10,11 The use of coronary CTA allows an objective measure of coronary artery disease (CAD) burden.10 In theory, once coronary anatomy has been determined by coronary CTA, any disparity in subsequent workup should not be the result of differences in presentation.
The goal of this investigation was to determine if there is a sex difference in rate of follow-up cardiac testing within 30 days after a coronary CTA performed for evaluation of ACS. We hypothesized that there should be no difference in cardiac test utilization between sexes found to have similar amounts of disease on coronary CTA.
We conducted a secondary analysis of a prospective cohort study of a convenience sample of low- to intermediate-risk patients presenting to the ED between 2005 and 2009 with potential ACS who received a coronary CTA as a part of their evaluation. As all coronary CTAs were standard of care and verbal consent was obtained consistent with the institutional review board (IRB) requirements for the initial study, the secondary analysis was approved with an IRB waiver of informed consent.
Study Setting and Population
The study was conducted at the Hospital of the University of Pennsylvania, an urban tertiary referral center with an annual census of 57,000 patients. The ED has 33 beds and a 17-bed observation unit where many patients are admitted for evaluation of chest pain.
Patients included in this study were a convenience sample of those who presented to the ED with a chief complaint consistent with ACS and received a coronary CTA in the ED or the observation unit. Patients at our institution receiving a coronary CTA in the ED have all been identified as low to intermediate risk by an attending physician, generally with a TIMI4,5 score of 0 to 2, however, not such low risk that they would otherwise be discharged without further testing. Patients with a serum creatinine level greater than 1.5 mg/dL and those with active asthma, pregnancy, or an iodinated contrast allergy were not eligible to receive a coronary CTA. Patients with external chest trauma or other obvious causes of chest pain were excluded.
Structured data collection was performed prospectively by the treating physician using a closed-question data instrument and included self-reported demographic characteristics, cardiac risk factors, chest pain characteristics, associated symptoms, medications and initial vital signs, electrocardiogram (ECG), treatment, and disposition in accordance with the standardized reporting guideline.12 For patients with a heart rate greater than 70 beats/min, intravenous or oral metoprolol was given. Coronary CTA was performed with predominantly 64-slice or 64-slice dual-source scanners (Siemens Medical Solutions, Malvern, PA). The protocol for coronary CTA image acquisition at our institution has been described elsewhere.13,14 All studies were read by attending radiologists with subspecialty cardiovascular imaging training, meeting American College of Cardiology/American Heart Association level III training guidelines. Study quality and technical limitations were included in the report, as well as any regions of the coronary arteries that could not be evaluated. All management decisions were left to the treating physicians.
Patients were contacted by trained research assistants 30 days after discharge by telephone and questioned about recurrent chest pain, nonfatal myocardial infarction, repeated hospitalization, and follow-up procedures such as stress tests, diagnostic cardiac catheterizations, PCIs, and surgery. Testing and procedures at our institution were then confirmed by medical record review. In the event that a subject could not be reached, a family member or friend was questioned. When neither the patient nor the secondary contact was able to be reached, medical records were reviewed. For patients for whom these methods failed to provide survival information, the Social Security Death Index (SSDI) was searched.
The primary outcome measure was whether the patient received a cardiac test (stress test, either pharmacologic or exercise, or diagnostic coronary catheterization) or was revascularized (PCI or CABG) within the 30-day follow-up period, stratified by coronary CTA results. Both inpatient and outpatient tests and procedures were included in analysis. All further testing decisions were made by the patient’s treating physician.
We also examined the number of coronary events (death, acute myocardial infarction [AMI], or revascularization procedure) at 30-day follow-up. Patients were analyzed in five prespecified groups according to the results of coronary CTA. Patients were stratified into categories by percent maximal stenosis: none, 1% to 24%, 25% to 49%, 50% to 69% and ≥70%. For patients with stenosis in multiple vessels, the greatest degree of stenosis in any artery was used. These cutoffs were defined based on previous literature.15,16
Data were analyzed using SAS v 9.2 (SAS Institute Inc., Cary, NC) and are presented as percent frequency occurrence with 95% confidence intervals (CIs). T-tests, chi-square, and Wilcoxon rank sum tests were used to compare male and female patients, as appropriate for continuous, categorical, and parametric or nonparametric data. We calculated the relative risk (RR) of receiving further testing by sex. Then to account for the possibility that one sex would have more diseased vessels than another, within each sex we stratified by maximal coronary CTA stenosis, as categorical variables described above, to give us an adjusted relative risk (aRR) of further cardiac testing. We conducted a multivariable regression model including age, sex, race, stenosis, and TIMI risk score to account for patient risk factors that may lead to differences in further testing. We then conducted a Mantel-Haenszel chi-square statistic to evaluate our hypothesis.
Our sample size was set a priori based on our enrolled cohort, which had over 600 females and 500 males. This enables us to have 80% power to detect true relative risks of 0.618 or 1.449 in male versus female subjects at an alpha of 0.05.
During the study period, 1,144 patients had coronary CTA performed and percent maximal stenosis recorded; 43 patients had testing performed but no stenosis recorded. Patients had a mean (±SD) age of 47.8 (±8.7) years, 55% were female, and 64% were black or African American. Male and female patients were generally similar with respect to many characteristics (Table 1); however, their racial distributions differed, and females were more likely to have a lower TIMI risk score and normal ECG. While the mean number of cardiac risk factors was similar between groups, females tended to have more hypertension, and males more hypercholesterolemia. We had 30-day follow up on 1,036 patients (90.5%), of whom 999 were able to be directly contacted; the others found by SSDI or record review.
|Characteristic||Women (n = 634)||Men (n = 510)|
|Age (±SD)||48.1 ± 8.5||47.1 ± 8.8|
|Black or African American||449 (70.8)||279 (54.7)|
|White||110 (17.3)||168 (32.9)|
|Cardiac risk factors (%),|
|Mean ± SD number of risk factors||1.7 ± 1.3||1.8 ± 1.4|
|Hypertension||313 (49.4)||203 (39.8)|
|Known CAD||6 (1.0)||6 (1.2)|
|Family history of CAD||115 (18.1)||102 (20.0)|
|Diabetes mellitus||97 (15.3)||66 (12.9)|
|Current tobacco use||209 (32.9)||208 (40.8)|
|Cocaine use in last week||38 (6.0)||62 (12.2)|
|Previous myocardial infarction||4 (0.6)||10 (1.9)|
|Previous stress test||158 (24.9)||132 (25.9)|
|Previous stress test abnormal||4 (0.6)||6 (1.2)|
|Previous cardiac catheterization||16 (2.5)||13 (2.6)|
|Previous cardiac catheterization abnormal||2 (0.3)||1 (0.2)|
|TIMI risk score (%)|
|0||403 (63.6)||278 (54.5)|
|2||46 (7.3)||60 (11.8)|
|3 or 4||5 (0.8)||13 (2.6)|
|Electrocardiographic findings (%)|
|Nonspecific||169 (26.7)||127 (24.9)|
|Early repolarization only||4 (0.6)||22 (4.3)|
|Abnormal not diagnostic||35 (5.5)||43 (8.4)|
|Ischemia, old-unchanged ECG||3 (0.5)||6 (1.2)|
|Ischemia, not known to be old||8 (1.3)||9 (1.8)|
|Suggestive of myocardial infarction||0 (0)||4 (0.7)|
|Normal coronary CTA results (%)||479 (75.6)||325 (63.7)|
Of the 1,144 patients included, 340 (29.7%) were found to have a stenosis of at least one vessel. Compared with women, more men had a stenosis of at least one vessel at coronary CTA (36.3% vs. 24.4%; odds ratio [OR] 1.76, 95% CI = 1.36 to 2.27). Coronary CTA results stratified by percent stenosis and sex are presented in Table 2. A total of 161 patients (14.1%) received at least one diagnostic test in the 30 days following their coronary CTAs, with 113 patients receiving their tests as inpatients during their index visits. Only one patient in the ≥70% maximal stenosis group did not have further testing as an in patient, and 10 patients in the 50% to 69% group did not have further testing. Overall, men were more likely to undergo further testing (RR = 1.51, 95% CI = 1.14 to 1.99).
|Maximal stenosis||n (%) with Further Testing||RR (95% CI)|
|Female (n = 591)||Male (n = 445)|
|None||23/442 (5.2)||20/282 (7.1)||1.27 (0.91–1.76)|
|1%–24%||1/24 (4.2)||2/27 (7.4)||1.28 (0.55–2.98)|
|25%–49%||12/66 (18.2)||12/67 (17.9)||0.99 (0.64–1.54)|
|50%–69%||23/43 (53.5)||27/42 (64.3)||0.99 (0.65–1.52)|
|70%–100%||15/16 (93.8)||26/27 (96.2)||0.93 (0.50–1.73)|
|Overall Mantel-Haenszel RR||1.14 (0.68–1.91)|
We constructed a generalized linear model incorporating age, race, sex, stenosis, and TIMI risk score to see if any of these other factors played a role in follow-up testing. Stenosis remained a significant factor (aRR = 1.51, 95% CI = 1.19 to 1.91), while male sex (aRR = 1.10, 95% CI = 0.86 to 1.42), age (aRR = 0.57, 95% CI = 0.29 to 1.17), race (aRR = 1.00, 95% CI = 0.77 to 1.30), and TIMI risk score (aRR = 0.80, 95% CI = 0.61 to 1.10) did not. Using a Mantel-Haenszel chi-square statistic, we stratified by percent stenosis and also found that men were not more likely to receive further testing within 30 days after coronary CTA compared to women (aRR = 1.14, 95% CI = 0.68 to 1.91, p = 0.98). Table 2 also shows rates of further testing (stress test or cardiac catheterization) at 30-day follow-up stratified by coronary CTA results and sex.
At 30-day follow-up we had follow up information on 1036 patients, and there was one cardiovascular death, nine revascularization procedures, two AMIs, and three patients who had both an AMI and a revascularization procedure. No coronary events occurred in individuals with less than a 50% maximal stenosis. Table 3 shows the number of cardiac events (death, myocardial infarction, or revascularization procedure) at 30-day follow-up according to coronary CTA results, as stratified by percent stenosis, sex, and whether further testing was performed; men and women were similar in each strata.
|Max Stenosis||0||1% to 24%||25% to 49%||50% to 69%||≥70%|
|Testing||0/23||0/20||0/1||0/2||0/12||0/12||2/23 (8.7)||1/27 (3.7)||4/15 (26.7)||7/26 (26.9)|
Given our missing data, we conducted sensitivity analyses. There was no significant difference in the proportion of men versus women for those missing stenosis data (23 women vs. 20 men, p = 0.76) or in those who received 30-day testing (74 women vs. 87 men, p = 0.34). In our analysis, we assumed that all missing patients did not receive follow-up testing. We conducted a sensitivity analysis and assumed that all missing patients had follow-up testing. Our results were still the same: males were more likely to get follow-up testing, but when taking stenosis group into account, there was again no significant relationship between sex and testing (aRR of all tests positive in missing = 1.19, 95% = CI 0.97 to 1.46; aRR of all tests negative in missing data = 1.15, 95% CI = 0.68 to 1.93).
Acute coronary syndrome represents a continuum of acute myocardial ischemia and is responsible for over 733,000 hospitalizations annually, nearly 50% of which occur in women.1 Previous studies have demonstrated that women are less likely to receive cardiovascular testing and interventions compared to men, even with ACS, despite the fact that cardiovascular disease is the leading cause of death in women in this country.17 While it is often cited that women with ischemic heart disease are more likely to present with atypical symptoms, several ED-based studies have shown that there are more similarities than differences between men and women.18,19 In a study conducted at our own institution, Chang et al.6 demonstrated that female patients presenting to the ED with potential ACS undergo fewer cardiac catheterizations and stress tests compared to male patients, even when adjusted for presenting complaint, history, ECG, and diagnosis. Possible explanations for the sex difference in testing include differences in presentation, physician bias, sex-specific patient preferences, or even an overtreatment of men.9
Sex differences associated with coronary heart disease and care have been studied extensively in a variety of settings; however, there are a number of challenges to identifying CAD in women that are important to discuss. Several studies have documented that women who receive a cardiac catheterization, the criterion standard for diagnosing and quantifying CAD, are less likely than their male counterparts to have CAD. In one study, fewer than half of women undergoing coronary angiography had obstructive CAD compared to more than 80% of men.2 Some have argued that this has led to lower rates of invasive testing and procedures in women.9 Another challenge to diagnosing CAD in women is that several of the most commonly used noninvasive diagnostic tests, such as stress testing, underperform in women.2,10 For example, the sensitivity and specificity of exercise ECG ranges from 60% to 70% for women compared to 80% for men.20 Reasons for the diminished accuracy in women include a lower prevalence of CAD for any single age group compared to men, a lower representation of women in earlier studies evaluating the diagnostic accuracy of tests, and a lower percentage of women who are able to reach target heart rate.2,21
Our study differs from previous studies evaluating sex bias in cardiac care in that we stratified patients by coronary CTA, a noninvasive test that has been shown to have excellent diagnostic accuracy when compared to cardiac catheterization.22 In contrast to stress testing, coronary CTA has been shown to have equal accuracy in men and women in excluding the presence of obstructive CAD.10,11 By stratifying patients according to coronary CTA results, we were able to control for any potential differences in presentation and the amount of disease present. This adds to the body of literature from our institution regarding sex disparities in cardiac care, as we have examined potential physician bias6 and patient preference.7 By adjusting for amount of disease, we are able to see that there are no differences by sex.
In our cohort of patients, men tended to have more extensive disease compared to women. Given that CAD develops at a later age in women compared to men, this finding was not surprising in our cohort with a mean age of 47.8 years. Because they had more extensive disease, it was also not surprising that, overall, men were more likely to receive further testing than women.
When results were adjusted for coronary CTA results, men were not more likely to receive further testing compared to women. This finding is in contrast to earlier studies examining follow-up care after initial abnormal cardiac test results. In their cohort of 3,975 patients referred for outpatient stress testing, Shaw et al.23 found that women with suspected coronary disease had fewer additional diagnostic tests than men after an initial abnormal noninvasive stress test result. In their cohort, 62% of men were referred for additional diagnostic testing compared to only 38% of women, even though the rates of positive stress test results were similar.
In their large nested case–control study of national Medicare patients, Lucas et al.24 explored the hypothesis that some of the observed sex differences in invasive cardiovascular procedures and testing documented in the literature could be due to different rates of entry into the “diagnostic cascade.” Their hypothesis was based on the assumption that cardiac catheterization is not typically the first event in the investigation of cardiac symptoms. They found that given the same exposure to the health care system and similar clinical conditions, non-African American men received more stress testing than African American men and all women. These differences were greatest among persons without indications for stress testing, and smallest for persons with indications suggesting that this disparity in testing could be partially due to “nonclinical” decision-making, and account for sex differences further down the diagnostic cascade.
There are several limitations within our study. First, our study population includes only patients who presented to the ED and were judged to be at low to medium risk for potential ACS. As a result, the majority of patients had normal coronary CTA (804 of 1,144), and there were only a small number of coronary events at 30-day follow-up. This study population does not necessarily reflect the patient population that most commonly undergoes stress testing and/or cardiac catheterization, nor does it reflect the population that has been studied before in the cardiovascular sex bias literature. Also, because our study sample is representative of our urban, predominantly black or African American ED population, it is unclear how generalizable the results are.
Second, because coronary CTA is a relatively new diagnostic tool, there is no consensus regarding if and what type of follow-up testing is necessary, especially when noncritical coronary disease is detected. Perhaps this is the reason why patients who had no stenosis detected on coronary CTA still went on to receive further cardiovascular testing. We realize that there are important differences between types of stress tests and cardiac catheterizations in degree of invasiveness and sensitivity and specificity for CAD detection; however, for the purpose of analysis they were grouped together because of the low rates for each type of testing in our cohort. Furthermore, all decisions regarding further testing were made on the part of the treating inpatient or outpatient physician. Perhaps physicians were unaware of coronary CTA results and patients who had critical stenosis were not evaluated further or vice versa. We feel that this is reflective of the “real-world” setting. We do believe, anecdotally, that as physicians became more comfortable with coronary CTA, further testing for low-grade lesions has decreased.
Last, variables were not analyzed that may have affected the rate of follow-up testing, including the patient’s insurance status, whether the patient has a primary care provider or cardiologist, and patient preference. Also, our analysis does not capture nonprocedural interventions such as medications or diet and exercise.
Our study suggests that female patients with potential acute coronary syndrome who undergo coronary computed tomography angiography as a part of their ED evaluation were not less likely than men to receive follow-up testing when stratified by coronary computed tomography angiography results.
- 1American Heart Association. Heart Disease and Stroke Statistics--2010 Update. Dallas, TX: American Heart Association, 2010.
- 16Incremental prognostic significance of left ventricular dysfunction to coronary artery disease detection by 64-detector row coronary computed tomographic angiography for the prediction of all-cause mortality: results from a two-centre study of 5330 patients. Eur Heart J. 2010; 31:1212–9., , , et al.