Author + information
- Received May 7, 2012
- Revision received December 31, 2012
- Accepted January 2, 2013
- Published online June 1, 2013.
- Gianluca Pontone, MD∗∗ (, )
- Daniele Andreini, MD∗,
- Antonio L. Bartorelli, MD∗,†,
- Erika Bertella, MD∗,
- Sarah Cortinovis, MD∗,
- Saima Mushtaq, MD∗,
- Claudia Foti, MS∗,
- Andrea Annoni, MD∗,
- Alberto Formenti, MD∗,
- Andrea Baggiano, MD∗,
- Edoardo Conte, MD∗,
- Francesca Bovis, MS∗,
- Fabrizio Veglia, MS∗,
- Giovanni Ballerini, MD∗,
- Cesare Fiorentini, MD∗,†,
- Piergiuseppe Agostoni, MD∗,† and
- Mauro Pepi, MD∗
- ∗Centro Cardiologico Monzino, IRCCS, Milan, Italy
- †Department of Cardiovascular Sciences, University of Milan, Milan, Italy
- ↵∗Reprint requests and correspondence:
Dr. Gianluca Pontone, Centro Cardiologico Monzino, IRCCS, Via C. Parea 4, 20138 Milan, Italy.
Objectives The aim of the study was to perform a comparison of the prognostic performance of computed tomography coronary angiography (CTA) and exercise electrocardiography (ex-ECG) in patients with suspected coronary artery disease (CAD).
Background CAD is a major cause of mortality and morbidity, and its management consumes a large proportion of the health care budget. Therefore, identification of patients at high risk of adverse events is crucial. Despite its limited accuracy, ex-ECG is the most commonly used noninvasive test in CAD evaluation. CTA was recently introduced as alternative test.
Methods We enrolled 681 patients (age 61.3 ± 10.4 years, 461 men) with atypical or typical angina and no history of CAD. All patients underwent ex-ECG and CTA and were followed for 44 ±12 months. The endpoints were all cardiac events, defined as nonfatal myocardial infarction, cardiac death, and revascularization, and hard cardiac events, defined as all cardiac events excluding revascularization.
Results ex-ECG and CTA were rated as positive in 419 (61%) and 274 (40%) of 681 patients, respectively. In univariate analysis, both ex-ECG and CTA were predictors of all cardiac events (hazard ratio [HR]: 2.09, 95% confidence interval [CI]: 1.5 to 2.8; p < 0.0001 and HR: 21.1, 95% CI: 14.6 to 30.5; p < 0.0001, respectively) and hard cardiac events (HR: 1.9, 95% CI: 1.1 to 3.2; p = 0.02 and HR: 6.8, 95% CI: 3.9 to 11.0; p < 0.0001, respectively), whereas in a multivariate analysis, CAD with ≥50% stenoses detected by CTA was the only independent predictor of hard cardiac events. Stratifying our population by ex-ECG and CTA findings, Kaplan-Meier curves showed that ex-ECG provides only a further risk stratification in the subset of patients with positive findings on CTA and a low to intermediate likelihood of CAD. Moreover, positive findings on CTA identify a shorter event-free period, regardless the ex-ECG findings for both all cardiac events and hard cardiac events, respectively.
Conclusions CTA may have a higher prognostic value compared with ex-ECG in patients with suspected CAD, mainly in those with a low to intermediate pre-test likelihood of CAD.
Coronary artery disease (CAD) is a major cause of mortality and morbidity, and its management consumes a large proportion of the health care budget. Therefore, identification of patients at high risk of adverse events is crucial. Exercise electrocardiography (ex-ECG) is the most commonly used noninvasive test for the diagnosis of CAD. However, ex-ECG stress has poor sensitivity and specificity (1). Indeed, ex-ECG may detect ischemia but not the presence of atherosclerosis in the absence of a flow-limiting stenosis. Computed tomography coronary angiography (CTA) was recently introduced as an anatomic imaging method for the evaluation of CAD. Several studies support the use of CTA to rule out the presence of CAD with high accuracy (2,3) and also for improving prognostic assessment above baseline risk factor evaluation (4–10). Retrospective comparative studies showed that CTA has a higher diagnostic accuracy compared with ex-ECG for detecting significant CAD (11,12). However, only a few studies have been conducted to compare the prognostic value of CTA and ex-ECG on a mid-term follow-up (13), whereas no data are available for long-term follow-up.
Therefore, the aim of the study was to compare the prognostic performance of CTA and ex-ECG in consecutive patients referred for suspected CAD with long-term follow-up.
We identified 3,750 consecutive patients referred to our hospital for diagnostic evaluation of atypical or typical angina between January 2005 and May 2010. We excluded from the present analysis patients with known CAD (n = 1,410) or known nonischemic cardiac disease (n = 680), pre-existing electrocardiographic changes or inability to perform ex-ECG with consequent inability to reach the target heart rate (n = 532), contraindications to contrast agents (n = 74), impaired renal function defined as creatinine clearance <60 ml/min (n = 112), inability to sustain a 15-s breath hold (n = 72), pregnancy (n = 11), and a resting heart rate ≥75 beats/min despite beta-blocker treatment or cardiac arrhythmias (n = 159). The remaining 700 patients were enrolled in the present study. They prospectively underwent both ex-ECG and CTA within 1 week of each other. Both tests were performed in addition to a standard clinical workup that was based on clinical evaluation and functional imaging stress test other than ex-ECG. ex-ECG and CTA findings were evaluated by different expert readers blinded to the clinical history of the patients, and similarly all events recorded during the follow-up were reviewed by an independent cardiologist who was blinded to both ex-ECG and CTA findings.
Our study complies with the Declaration of Helsinki. The study was approved by the local ethics committee. All patients gave written informed consent.
A structured interview was performed and a clinical history was obtained, and the following cardiac risk factors were assessed in all patients before CTA: 1) hypertension (blood pressure >140/90 mm Hg or use of antihypertensive agents) (14); 2) smoking (currently or previously); 3) hyperlipidemia (low-density lipoprotein cholesterol >140 mg/dl) (15); 4) diabetes mellitus (fasting glucose level >110 mg/dl or the need for insulin or oral antidiabetes drugs) (16); 5) family history of CAD in first-degree relatives (17); 6) home use of antianginal drugs; and 7) symptoms (atypical or typical angina) (18). According to the Diamond criteria (18), the study population was classified as having a pre-test low to intermediate (<70%) or high (70%) likelihood of CAD.
Each patient performed a cycle ergometer-graded exercise test. The stress test response was considered positive in case of: 1) horizontal or downsloping ST-segment depression >0.1 mV measured at 80 or 60 ms after the J-point during exercise or recovery; 2) upsloping ST-segment depression of 0.15 mV at 80 ms after the J-point; and 3) ST-segment elevation >0.1 mV measured at 80 or 60 ms after the J-point during exercise or recovery.
CTA protocol and dataset evaluation
CTA was performed between 1 and 3 days (average, 2.2 ± 0.4 days) after ex-ECG. In all patients with resting heart rate >65 beats/min before CTA, metoprolol was intravenously administered with a titration dose up to 15 mg to achieve a target heart rate ≤65 beats/min.
CTA was performed with a 64-slice scanner (GE Healthcare, Milwaukee, Wisconsin). Retrospective electrocardiographic triggering with electrocardiographic pulsing technique was used before September 1, 2007, whereas prospective electrocardiographic triggering (SnapShot Pulse, GE Healthcare) was used in patients who were enrolled after September 1, 2007, and in whom a heart rate ≤65 beats/min was reached.
Image CTA datasets were transferred to a dedicated workstation and analyzed with cardiac software (Advantage Workstation and CardioQ3 Package, GE Healthcare) by 2 expert readers (G.P. and D.A.) blinded to the ex-ECG findings. For any disagreement in data analysis between the 2 readers, consensus agreement was achieved involving a third expert reader.
Coronary artery segments were classified according to the 16-segment American Heart Association classification (19), including in our analysis all segments with a diameter of at least 1.5 mm at their origin. Each segment was classified as assessable or not assessable in the presence of severe artifacts impairing accurate evaluation. Patients with proximal or mid segment or >3 segments classified as not assessable were rated as positive (5). For each assessable segment, the presence of coronary plaque was defined as a structure >1 mm2 within and/or adjacent to the artery lumen (4). Then, each patient was classified in a patient-based model as having CAD <50% (at least 1 segment with a stenosis <50% but no segment with a stenosis ≥50%), CAD ≥50% (at least 1 segment with a stenosis ≥50%) and in a vessel-based model as having left main coronary artery, 1-, 2-, or 3-vessel disease using a stenosis threshold ≥50%. Finally, we calculated the segment involved score measured as the sum of coronary segments (minimum = 0, maximum = 16) showing the presence of plaque irrespective of stenosis severity and the segment stenoses score calculated as the sum of each segment score as defined above (total score ranging from 0 to 48) (20).
Patient follow-up was performed by checking the medical records or by phone interview by researchers unaware of the patients’ CTA results. Events were defined as follows: nonfatal myocardial infarction (typical chest pain with elevated cardiac enzyme levels and typical ST-segment changes on the electrocardiogram) (21), late revascularization defined as elective revascularization 60 days after CTA and cardiac death (death caused by acute myocardial infarction, ventricular arrhythmias, or refractory heart failure). We defined hard cardiac events as a combined endpoint of nonfatal myocardial infarction and cardiac death. All cardiac events were defined as hard cardiac events plus revascularization of any sort. In case of multiple events in a given patient, the first event was included in the analysis.
According to CTA and ex-ECG findings, the study population was divided in 4 groups: group A (CAD <50% and negative ex-ECG), group B (CAD <50% and positive ex-ECG), group C (CAD ≥50% and negative ex-ECG), and group D (CAD ≥50% and positive ex-ECG). Categorical baseline characteristics were expressed as numbers and percentages, whereas continuous variables were expressed as mean ± SD. To identify the association between ex-ECG findings, CTA variables, and outcomes, Cox regression analysis was used. First, univariate analysis of clinical characteristics and CTA and ex-ECG findings was performed to identify potential predictors. Hazard ratios were calculated with 95% confidence intervals as an estimate of the risk associated with a particular variable. To determine independent predictors of the composite endpoints, multivariate analysis of variables with p ≤ 0.05 in univariate analysis was performed and corrected for the following baseline characteristics: male sex, age, diabetes, hypercholesterolemia, hypertension, family history of CAD, and smoking. Finally, to avoid overfitting and multicollinearity issues, we developed a model adjusted for coronary artery classification in 1-, 2-, and 3-vessel disease and left main coronary artery with CAD ≥50%.
Cumulative event rates for each group were obtained with the Kaplan-Meier method for all cardiac events and hard cardiac events and compared with the Wilcoxon log-rank test in all study population and in patients with a low to intermediate and high pre-test likelihood of CAD. Statistical analyses were performed using SPSS software, version 17.0 (SPSS Inc., Chicago, Illinois) and the SAS software version 6.12 (SAS Institute Inc., Cary, North Carolina). A p value <0.05 was considered significant.
The clinical characteristics of the study patients are given in Table 1. Of 700 patients, 10 were excluded because a heart rate <75 beats/min was not reached and 9 patients declined to participate in the follow-up. Therefore, 681 patients (mean age, 61.3 ± 10.4 years; 461 men) were included in our analysis.
Mean follow-up duration was 44 ± 12 months. A total of 37 patients (0, 0, 10, and 27 in groups A, B, C, and D, respectively) underwent early elective revascularizations and therefore were excluded from the analysis. During follow-up, all cardiac event and hard cardiac event endpoints were reached in 251 (37%) and 73 (11%) of patients, with an event-free survival time of 245 ± 358 days and 608 ± 432 days, respectively.
Based on CTA findings, 274 (40%) patients had CAD ≥50% with a segment involved score and segment stenoses score of 1.8 ± 2.2 and 4.7 ± 6.6, respectively. The overall agreement between ex-ECG and CTA was 55% (378 of 681 patients). Indeed, CTA and ex-ECG results showed agreement for the absence 195 (28%) patients (group A) and presence 183 (27%) patients (group D) of significant CAD in and 224 (33%) patients, respectively. In contrast, with CAD <50% by CTA had positive ex-ECG findings (group B) and 79 (12%) patients with CAD 50% had negative ex-ECG findings (group C). All cardiac events and hard cardiac events in groups A, B, C, and D were, respectively, 12 (7%), 22 (10%), 53 (67%), and 164 (81%) and 7 (4%), 8 (4%), 13 (16%), and 45 (23%), with an annual rate of 2%, 3%, 20%, and 24%, and 1%, 1/%, 4%, and 6.3%, respectively. All cardiac and hard cardiac events had a lower incidence in groups A and B versus groups C and D (p < 0.01). Of note, no difference in terms of events was observed between groups A and B, whereas group D showed a higher incidence of all cardiac events (p < 0.01). Moreover, positive CTA findings identify a shorter event-free period versus negative CTA findings regardless the ex-ECG findings for both all cardiac events (p < 0.01) and hard cardiac events (p < 0.05).
At baseline, the following imaging stress tests included in the workup decision making such as stress echocardiography, stress nuclear perfusion imaging, and stress cardiac magnetic resonance were performed in 337 (49%), 209 (31%), and 135 (20%) patients, respectively. These tests were censored as negative in 313 of 407 patients without CAD on CTA and positive in 157 of 274 patients with CAD ≥50%, with a mean overall agreement significantly higher than the agreement between ex-ECG and CTA (69% vs. 55%, p < 0.01). The rate of revascularization was higher in patients with abnormal versus normal imaging stress test findings (66% vs. 11%, p < 0.01) because these tests were used as a gatekeeper to invasive coronary angiography and potential revascularization. Hard cardiac events occurred in 23 of 430 and in 50 of 251 patients with normal and abnormal imaging stress test findings, respectively, with an annualized hard event rate of 1.4% and 5.4%, respectively (p < 0.001). Of note, in the subset of patients with normal imaging stress test results, the annualized hard cardiac event rate increased from 0.8% to 3% when associated with negative and positive CTA findings, respectively (p = 0.0012).
Table 2 summarizes the univariate and multivariate analyses of the clinical characteristics and CTA and ex-ECG results that were used for event prediction. In the multivariate analysis, the presence of CAD ≥50% on CTA remained the only independent predictor for all and hard cardiac events.
Kaplan-Meier curves showed that CTA has a better prognostic performance compared with ex-ECG (Fig. 1). From the combined analysis of ex-ECG and CTA findings, it was found that CTA always provided a prognostic stratification, whereas ex-ECG predicted all and hard cardiac events only in the subset of patients with CAD ≥50% on CTA (Fig. 2), particularly in those at high risk in whom a wrong CAD classification on CTA may be possible (Fig. 3).
The main findings of this study are the following: 1) CTA shows a better prognostic performance compared with ex-ECG; 2) evaluation of coronary anatomy with CTA may be the first diagnostic tool needed for prognostic stratification of patients with a low to intermediate pre-test likelihood of CAD, whereas ex-ECG may be more appropriate for further prognostic stratification in the subset of patients with CAD ≥50% on CTA; and 3) positive CTA findings identify a shorter event-free survival time regardless of the presence of ischemia at ex-ECG.
In the management of patients with suspected CAD, the prognostic stratification plays a crucial role beyond the simple diagnosis of coronary artery stenoses. Indeed, the occurrence of adverse events determines morbidity and mortality and influences the overall health expenditure. Until few years ago, diagnosis and prognostic evaluation of patients with suspected CAD were made with functional stress tests only in the majority of patients. Nowadays, ex-ECG is still regarded as the first-line functional examination in patients with chest pain, and it is the most widely used test in clinical practice for CAD assessment and for refining patient prognosis. Indeed, ST-segment changes, exercise capacity, chronotropic response, heart rate recovery, and ventricular ectopy during effort are all well-established independent risk factors for all-cause cardiovascular mortality (22). However, the prognostic value of ex-ECG is challenged by the inability to reflect the impact on prognosis of an atherosclerotic plaque when it does not impair coronary flow (22). In this regard, the recent introduction of CTA allows a noninvasive anatomic evaluation of symptomatic patients. Robust data regarding its accuracy in the detection of significant CAD have been reported in the literature (2,3). More recently, a potential prognostic role of this diagnostic tool has also emerged (4–10). Pundziute et al. (5), using CTA, evaluated a cohort of 104 patients who were followed for 16 months and found that obstructive CAD detection, the presence of coronary plaques irrespective of stenosis severity, the number of coronary segments with plaques, and the number of coronary arteries with mixed plaques were all independent predictors of cardiac events. This was the first report suggesting that the atherosclerotic burden of coronary arteries may influence patient prognosis beyond stenosis severity. In a later study, Min et al. (4) demonstrated in a larger cohort of patients that Duke and segment stenoses scores were significant predictors of all-cause mortality. Regarding the comparison of ex-ECG and CTA, Dewey et al. (11) showed a higher sensitivity and specificity of CTA versus ex-ECG for the detection of significant CAD, whereas in a previous study, we demonstrated a complementary role of the 2 tests in patients with equivocal clinical presentation (12). More recently, Dedic et al. (13) compared the prognostic performance of ex-ECG with that of CTA in 424 consecutive patients followed for 2.6 years. Similar to our study, they demonstrated that the presence of obstructive CAD on CTA showed incremental value beyond exercise testing in the prediction of future adverse events. However, compared with our study, the sample size was smaller, the duration of follow-up was shorter, ex-ECG findings were not used for clinical decisions, no data regarding combined endpoint without revascularization were reported, and nondiagnostic ex-ECG results were not excluded from the analysis.
These findings may be of clinical relevance. First, the number of vessels with coronary stenoses >50% is an independent predictor of adverse events. Second, ex-ECG retains a prognostic value in patients with positive CTA findings. Previous clinical observations may explain our findings. Indeed, it has been shown that adverse events are not due to flow-limiting coronary stenoses only, the presence of which is well detected by ex-ECG, but also due to the atherosclerotic plaque burden, apart from the degree of coronary flow reduction (23). This suggests that an integrated approach of anatomic and functional tests may be beneficial when assessing patients with suspected CAD. In this regard, our study shows that CTA works well in patients with a low to intermediate pre-test likelihood of CAD, whereas its prognostic stratification becomes weaker when the patient’s cardiovascular risk increases. The reduced performance of CTA in the latter case is likely due to an increased rate of inaccurate classification of coronary stenoses in high-risk patients. Indeed, decreased diagnostic accuracy of CTA with the increase of pre-test likelihood of CAD was previously reported (3,24). This suggests that in this patient subset, use of ex-ECG is definitely needed for further stratification. Whether other functional tests provide better prognostic stratifications compared with ex-ECG is beyond the scope of this study. In this regard, it has been shown that stress-rest myocardial perfusion imaging is superior to ex-ECG for the detection of myocardial ischemia and provides independent and incremental information in predicting future cardiac events (25). In agreement with these findings, van Werkhoven et al. (8) showed that the combined use of CTA and a nuclear stress test not only provides information with respect to the presence, extent, and composition of atherosclerosis, but, more importantly, improves risk stratification compared with myocardial perfusion imaging alone. In our study, despite that this point is beyond the aim of the study, a subanalysis of imaging stress tests included in the diagnostic workup decision making was performed, confirming that a positive CTA finding is associated with a higher rate of hard cardiac events in the subset of patients without ischemia on imaging stress tests, as previously demonstrated (8). Further prospective and blinded studies comparing an integrated approach using CTA and different functional tests are needed to show which diagnostic algorithm provides the best prognostic stratification. However, when such a combined strategy that includes CTA is proposed as a systematic diagnostic approach, the issues of costs, cumulative radiation exposure, and nephrotoxicity have to be considered. First, despite the fact that the cost of CTA is higher than that of ex-ECG, the evaluation of CAD by CTA followed by ex-ECG may lower overall health care costs primarily by decreasing the rate of false-positive results on ex-ECG. Moreover, due to its better prognostic value, patients evaluated by CTA may have fewer CAD-related episodes of care and consequently lower overall CAD-related costs. Second, the radiation exposure associated with CTA is an issue of concern. Although radiation exposure may differ in relation to different CTA techniques and protocols, a radiation dose up to 29 mSv may be reached with the first generation of 64-slice scanners (26). This dose level may raise concern regarding radiation-induced carcinogenesis. To put these radiation levels in context, it has been estimated that a coronary CTA angiogram with an effective dose of 10 mSv has a risk of inducing a fatal cancer in 1 in 2,000 patients (27). In this regard, the American Heart Association published a statement in 2006 in which they advised that a 10-mSv threshold should not be exceeded. Moreover, submillisievert exposure was recently reached with the latest scanner generation (28), suggesting a potential extensive and safer use of this technique in association with a nonradiation diagnostic tool such as ex-ECG for risk stratification. Third, the risk of nephrotoxicity in patients with normal renal function is negligible. Indeed, no cases of renal failure were observed in our population.
The results of our study suggest that patients with suspected CAD may be divided into 3 groups with different management and follow-up strategies: 1) patients with normal coronary anatomy in which short- or mid-term follow-up is not required; 2) patients without flow-limiting coronary atherosclerosis in whom drug treatment, aggressive risk factor modification, and mid-term follow-up with frequent ex-ECG are needed; and 3) patients with flow-limiting stenoses in whom further short-term evaluation with functional stress testing is highly advisable to assess the need and choose the type of myocardial revascularization. However, this proposed clinical strategy needs to be evaluated in a dedicated multicenter study.
First, this was a single-center study. A second limitation was the limited number of hard cardiac events due to the stable condition of our population. Third, in the ex-ECG prognostic evaluation, we did not emphasize the role of nonelectrocardiographic parameters such as functional capacity, chronotropic response, heart rate recovery, and ventricular ectopy, which may add further prognostic information beyond that provided by ST-segment changes. However, these parameters are rarely used in clinical practice for clinical decision making.
CTA may show a higher prognostic value compared with ex-ECG in patients with suspected CAD, mainly in those with a low to intermediate pre-test likelihood of CAD. In this clinical setting, CTA may be a better diagnostic tool and may play a key role in prognostic stratification in a combined anatomic and functional assessment workflow.
All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary artery disease
- confidence interval
- computed tomography coronary angiography
- exercise electrocardiography
- hazard ratio
- Received May 7, 2012.
- Revision received December 31, 2012.
- Accepted January 2, 2013.
- American College of Cardiology Foundation
- Budoff M.J.,
- Dowe D.,
- Jollis J.G.,
- et al.
- Min J.K.,
- Shaw L.J.,
- Devereux R.B.,
- et al.
- Pundziute G.,
- Schuijf J.D.,
- Jukema J.W.,
- et al.
- Carrigan T.P.,
- Nair D.,
- Schoenhagen P.,
- et al.
- van Werkhoven J.M.,
- Schuijf J.D.,
- Gaemperli O.,
- et al.
- Chow B.J.,
- Wells G.A.,
- Chen L.,
- et al.
- Dewey M.,
- Dübel H.P.,
- Schink T.,
- Baumann G.,
- Hamm B.
- ↵(1997) Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 20:1183–1197.
- Taylor A.J.,
- Bindeman J.,
- Feuerstein I.,
- Cao F.,
- Brazaitis M.,
- O’Malley P.G.
- ↵(2000) Myocardial infarction redefined—a consensus document of The Joint European Society of Cardiology/American College of Cardiology Committee for the redefinition of myocardial infarction. Eur Heart J 21:1502–1513.
- Hoffmann U.,
- Moselewski F.,
- Cury R.C.,
- et al.
- Diaz L.A.,
- Brunken R.C.,
- Blackstone E.H.,
- Snader C.E.,
- Lauer M.S.
- Coles D.R.,
- Smail M.A.,
- Negus I.S.,
- et al.