Exercise Electrocardiogram Neither Predicts Nor Excludes Coronary Artery Disease in Women with Low to Intermediate Risk

Abstract

Aim:

The value of exercise electrocardiogram (ExECG) in symptomatic female patients with low to intermediate risk for significant coronary artery disease (CAD) has been under debate for many years, and nondiagnostic or even erroneous test results are frequently encountered. Cardiac-CT may be more appropriate to exclude CAD in women. This study compares the results of ExECGs with those of cardiac-CTs, performed within a time frame of 1 month in an all-comers female chest pain population.

Patients and Methods:

Five hundred fifty-one consecutive female patients from a patient registry were included. ExECGs were negative in 324 (59%), positive in 14 (3%), and nondiagnostic in 213 (39%) patients. CAD was revealed by cardiac-CT in 57% of the women with negative ExECG. No signs of CAD were present on cardiac-CT in 64% of the women with a positive ExECG. Cardiac-CT showed presence of CAD in 268/551 (49%) patients, of whom 56/268 (21%) was diagnosed with ≥50% stenosis. The ExECG of the latter group was negative in 26 (46%), inconclusive in 29 (52%), and positive in 1 (2%). Considering ≥50% stenosis at cardiac-CT as the reference, sensitivity, specificity, PPV, and NPV of ExECG for the present population were 3.7%, 95.7%, 7.1%, and 91.7%, respectively. Similar diagnostic performance was calculated when considering ≥70% stenosis at cardiac-CT as the reference.

Conclusion:

ExECG failed to detect CAD in more than half of this cohort and in almost half of women with >50% stenosis at cardiac-CT. Importantly, no CAD was detected by cardiac-CT in 64% of women with a positive ExECG. ExECG is therefore questionable as a diagnostic strategy in women with low-to-intermediate risk of CAD, although prospective studies are warranted to determine whether replacing ExECG by cardiac-CT provides better prognoses.

Introduction

Despite many technological advancements in the past decades, exercise electrocardiogram (ExECG) is still widely used to evaluate coronary artery disease (CAD) in patients presenting with chest pain. To date, ExECG remains the preferred study for functionally capable women with low-to-intermediate risk for CAD when an interpretable resting ECG is present.¹ The simplicity, wide availability, and cost-effectiveness of ExECG are solid reasons to perform such tests in chest pain patients.

However, although ExECG is a good diagnostic method in certain categories of patients, it has its limitations. The ESC guideline has indeed allowed noninvasive imaging techniques a more prominent place in the diagnostic process of CAD.² For example, ExECG is known to be less valuable in female compared to male patients, and ExECG tests often result in a nondiagnostic outcome for a diversity of reasons. Based on a number of reports, both sensitivity and specificity^3,4 of the test have been questioned in symptomatic female patients, although this has been shown to improve when additional parameters such as exercise capacity and the Duke Treadmill Score⁵ are combined with the ExECG results.⁴

Yet, the interpretation of an ExECG may remain problematic in a subset of female patients with a low pretest likelihood for significant coronary artery stenosis and atypical angina, or in patients who are unable to deliver the required exercise.² Ruling out CAD in patients with low-to-intermediate risk of significant CAD may have several clinical implications. These patients have an excellent cardiovascular prognosis and unnecessary prescriptions of aspirin or statins can be detected, while concerns of heart disease can be better counseled.⁶

Recent developments in noninvasive imaging procedures in cardiology may allow for a more accurate first-line diagnosis of CAD, at least in this subset of patients. Coronary computed tomography angiography (CCTA) and/or calcium scoring (CaSc) are well-established methods for the exclusion or grading of CAD. A negative CCTA is associated with a favorable long-term prognosis, whereas obstructive as well as nonobstructive plaques on CCTA have been shown to correlate well with future cardiac events.⁷ Using modern and dedicated scanning equipment, optimized scan protocols, and proper patient selection, both imaging studies can be performed with an acceptable radiation burden.^8,9

A direct comparison in female chest pain patients with a low-to-intermediate risk for significant CAD between the outcome of ExECG and CCTA in a regular clinical population is lacking in literature. The objective of this analysis was to compare the ExECG results with the CCTA diagnosis in female patients with a low-to-intermediate probability for CAD in an all-comers chest pain population.

Patients and Methods

Cardiac-CT register

From January 2012 through January 2015, data from 3145 patients (1301 males and 1844 females) referred for cardiac-CT, both CaSc and CCTA, were prospectively included in a database. Patients were referred from the outpatient clinic after a diagnostic workup for chest pain or after previous evaluation at the emergency department. In the latter, acute coronary syndrome had been excluded and troponin measurements were negative for all patients. Of the 1844 female patients, 551 (30%) underwent an exercise test as well as a CCTA and were selected for this analysis. Pregnant and lactating patients were considered not eligible for CCTA, as were patients allergic to intravenous radiocontrast, when not adequately pretreated with steroids.

All patients gave written informed consent for the use of their anonymous data for scientific purposes. Besides the standard imaging protocol, no additional measurements or actions affecting the patient were performed. Therefore approval of the local ethics committee for this study was not necessary since the study does not fall within the scope of the Dutch Medical Research Involving Human Subjects Act (section 1.b WMO, 26th February 1998).

The following data were obtained for all patients: age, risk factors (hypertension, hypercholesterolemia, diabetes mellitus, positive family history of CAD, and smoking history), weight, length, body mass index (BMI), and estimated glomerular filtration rate. Baseline medication use and postimaging changes in therapy were both prospectively documented in the database, as was referral for additional myocardial perfusion scintigraphy or coronary angiography (CAG).

ExECG test variables

All ExECGs performed within 1 month before or after CCTA were retrieved from the patients' records and the following data were extracted: date of ExECG, exercise time, maximal load in Watt, rest heart rate (HR), maximal HR, percentage of predicted maximal HR, blood pressure, main reason for termination of the ExECG, metabolic equivalents (METs), complaints, ST segment changes, and cardiac arrhythmias.

ExECG and interpretation

ExECG was performed on an exercise bicycle with an initial load of 20 Watt and increments of 10–20 Watt each minute. The ECG was recorded and monitored continuously, and blood pressure, HR, presence of chest pain, other complaints, and electrocardiography changes in each stage and recovery phase (3–5 minutes after peak exercise) were recorded. Indications for termination of the test included ST segment elevation greater than 2.0 mm (0.2 mV) in two or more contiguous precordial leads or adjacent limb leads, systolic blood pressure drop >10 mmHg from baseline, severe hypertension (>250 mmHg systolic or >115 mmHg diastolic), significant arrhythmia (ventricular tachycardia or atrial fibrillation), moderate-to-severe angina, cyanosis or pallor, symptoms of dizziness/near-syncope or exhaustion, or muscle acidification.

The clinician's interpretation of the exercise test was documented in the electronic patient record and used for this analysis. The results of the ExECG were categorized as positive, nondiagnostic, or negative. A positive ExECG was defined as ≥1 mm ST segment depression, downsloping or horizontal, present at more than 60 milliseconds after the J-point. A negative ExECG was defined as absence of ST segment depression. A nondiagnostic ExECG was defined as failure to reach 85% of the predicted maximal HR in accordance to age or when <1 mm ST segment depression was noticed on the ECG.

CaSc and CCTA acquisition and interpretation

CaSc and CCTA were obtained by a dual source flying focal spot 2 × 64 slice (resulting in 2 × 128 slices) Somatom Definition Flash (Siemens Medical Systems, Erlangen, Germany) using a gantry rotation of 280 milliseconds and temporal resolution of 75 milliseconds. The CaSc scan was obtained using a tube voltage of 120 kV, with 3 mm slice thickness, in high pitch (3.4) flash mode or prospective mode (step and shoot) in best diastolic phase, whereas the tube current was automatically determined by the system using CARE Dose 4D (Siemens Medical Systems). The dataset was filtered using a B35f HeartView medium filter. Areas with Hounsfield units >130 were considered to contain calcium, manually assigned to the appropriate coronary artery and added to the Agatston score.¹⁰

Subsequently, CCTA was performed using a tube voltage between 80 and 120 kV, which was automatically determined by the system, as was tube current. The detector collimation was 2 × 64 mm. Next, a prospective ECG-triggered acquisition with a pitch of 3.4 in flash spiral mode was performed in best diastolic phase, an acquisition in prospective mode (step and shoot), or an acquisition in retrospective mode throughout the cardiac cycle. The images were reconstructed with a slice thickness of 0.6 mm with an increment of 0.3 mm for all flash acquisitions and an increment of 0.4 mm for prospective or retrospective acquisitions. Details of the acquisition protocols were previously described.^11,12

On a patient basis, the scans were scored for the presence or absence of coronary disease. Hyperdense, hypodense, or mixed structures >1 mm² in and/or adjacent to the coronary artery wall were considered coronary plaque. Absence of CAD was defined as CaSc 0 and no evidence of calcified or noncalcified plaques. Presence of CAD was defined as CaSc >0 and/or presence of calcified or noncalcified plaques.

Comparison of ExECG and cardiac-CT results

Data from the ExECG and CCTA were pooled and compared. Discordances between presence of CAD and significant coronary artery lumen stenosis on cardiac-CT and ExECG results were scored for all included patients.

Estimation of radiation dose of cardiac-CT

After each procedure, the scanner automatically generated the delivered radiation dose to the patient, reflected as the dose length product (DLP). The effective dose of the scan was estimated by using a k-factor of 0.014 mSv/(mGy·cm) as proposed by the European Working Group for Guidelines on Quality Criteria in CT. In this study, the total dose of topogram, test bolus, CaSc, and CCTA was used to estimate the effective radiation dose for each patient.

Follow-up

Follow-up data were retrieved from the patient records for two groups of particular interest, including the group of female patients, which was diagnosed with coronary artery stenosis of >50% using CCTA. Follow-up data were also retrieved for the group of female patients with negative ExECGs. The follow-up period extends from the day of the performed test (CCTA or ExECG) to the date of analysis.

Statistics

SPSS version 20.0 (IBM Co., Armonk, NY) was used for analysis of data. Normally distributed continuous variables were expressed as mean and standard deviation (SD) in case of normality (tested using the Kolmogorov–Smirnov test). The relationship between CAD (diagnosed with cardiac-CT) and patient variables and the relationship between normal and abnormal ExECG and patient variables were tested using logistic regression in case of dichotomous dependent variables and linear regression in case of continuous dependent variables. Multivariable analyses were performed on variables that resulted in p < 0.20 in univariate analyses. p-values <0.05 were considered significant.

The Duke Clinical Score¹³ was not analyzed in the univariate and multivariate analyses since gender, age, DM, and hypercholesterolemia are parameters in this particular scoring method.

Results

Patients

Five hundred fifty-one female patients, all of whom were subjected to ExECG within 1 month before or after the CCTA, were included in the study. Baseline characteristics of these women were (mean ± SD) as follows: age 58 ± 10 years, length 168 ± 7 cm, weight 75 ± 14 kg, and BMI 26.6 ± 4.9. Cardiovascular risk factors are listed in Table 1.

Table 1.

Patient Characteristics and Risk Factors

n^a	551
Age, years (mean ± SD)	58 ± 10
Length, cm (mean ± SD)	168 ± 7
Weight, kg (mean ± SD)	75 ± 14
Body mass index, kg/m² (mean ± SD)	26.6 ± 4.9
Duke clinical score (mean ± SD, n = 517)^b	23.4 ± 19.1
Risk factors
Positive family history	295 (54%)
Smoking	126 (23%)
Diabetes Mellitus	39 (7%)
Hypercholesterolemia	157 (29%)
Hypertension	186 (34%)

Number of included women.

Data missing in 34 women.

In 34 women, the Duke Clinical Score was missing from the patient records, whereas in the other women, this score was 23.4 ± 19.1. In the evaluated cohort, the prevalence of positive family history, smoking, diabetes, hypercholesterolemia, and hypertension was in 295/551 (54%), 126/551 (23%), 39/551 (7%), 157/551 (29%), and 186/551 (34%) patients, respectively. Baseline low-dose aspirin treatment, baseline statin treatment, and treatment with a combination of the two medications were reported in 225/551 (41%), 179/551 (33%), and 144/551 (26%) patients, respectively.

ExECG results

After analysis of the 551 ExECGs, the characteristics of the exercise tests proved to be as follows: achieved exercise load 133 ± 38 W (range 40–300 W), METs 5.4 ± 1.6 (range 1.6–12.8), and duration of the exercise of 6:43 ± 1:54 min (range 2:20–15:45). In 308/551 (55%) patients, the main reason for termination of the exercise was exhaustion, in 122/551 (22%) dyspnea, and in 68/551 (12%) muscle acidification. The remaining 11% had various reasons for termination of the test such as sufficient exercise load and HR response, while three women showed significant ST segment depression, which led directly to termination of the exercise test.

The rest HR in the tested cohort was 76 ± 14 and the maximum HR was 142 ± 22 beats per minute. The rest diastolic and rest systolic blood pressure were 80 ± 12 and 134 ± 22 mmHg, whereas the stress diastolic and stress systolic blood pressure were 82 ± 14 and 184 ± 28 mmHg, respectively. In five patients, the rest diastolic RR and rest systolic RR before ExECG were missing from the patient records.

ExECGs were categorized as negative in 324/551 (59%) patients, positive in 14/551 (3%) patients, and nondiagnostic in 213/551 (39%) patients. Of the latter, 190/213 (89%) failed to achieve >85% of the maximum of the expected HR and in 23/213 (11%), ST segment depression was detected, but did not meet the criteria of ≥1 mm ST segment depression.

Cardiac-CT results

In 532 women, both CaSc and CCTA were obtained, whereas in 19 women, only a CaSc was performed. The Agatston score in the latter group was 948 ± 631 (mean ± SD; range 0–1853). From these women, two patients were referred for CAG and 14 women were redirected to ¹³N-ammonia myocardial perfusion PET/CT. CCTA was cancelled in one young woman with an Agatston score of zero, atypical chest complaints, and very low pretest probability, to minimize radiation exposure. For similar reasons, CCTA was omitted in one relatively young woman with an Agatston score of 25, but irregular HR due to atrial fibrillation. Another patient with an Agatston score of 1853 refused further imaging procedures due to severe claustrophobia.

Of all performed CCTAs, 388 (73%) were acquired in high-pitch flash mode, whereas 136 (26%) were obtained in prospective mode and 8 (2%) in retrospective mode.

In the entire cohort, the Agatston score was 97 ± 269. CAD was diagnosed in 268/551 (49%) patients, whereas in 278/551 (51%) patients, no CAD was detected. Coronary artery lumen stenosis of >50% was demonstrated in 56/268 (21%) of the patients who were diagnosed with CAD, whereas 194/268 (72%) of the patients were diagnosed with <50% stenosis. In 19 patients (3% of the total cohort), who were only subjected to CaSc, data on significance of stenosis were missing.

Alterations in patient management based on the results of cardiac-CT were observed in 267/551 (49%) patients, while in 281/551 (51%) patients, patient management remained unchanged. In 3/551 (<1%) of the patients, these data were missing in the database. Alterations in treatment regime typically comprised changes in medication, including changes of more than one medicine in a single patient. In 46/551 (8%) patients, low-dose aspirins were started, in 103/551 (19%) patients, statins were started, in 25/551 (5%) patients, β-blockers were started, and in 11/551 (2%) patients, angiotensin-converting enzyme (ACE) inhibitors were initiated. Low-dose aspirins were discontinued in 101/551 (18%) patients, statins were stopped in 65/551 (12%) patients, β-blockers in 67/551 (12%) patients, and ACE inhibitors in 6/551 (1%).

In case of a high Agatston score, other alterations in treatment regime included referral for CAG or ¹³N-ammonia myocardial perfusion PET/CT, as mentioned above.

Radiation dose of cardiac-CT

The mean radiation dose was 2.1 ± 1.9 mSv (range 0.31–16.3 mSv). The mean ± SD radiation doses for the high-pitch flash CCTAs, prospective mode CCTAs, retrospective mode CCTAs, and CaSc scans were 1.4 ± 0.6, 4.0 ± 2.5, 6.0 ± 4.4, and 0.5 ± 0.2 mSv, respectively.

Comparison of ExECG and cardiac-CT

Table 2 displays a comparison between the ExECG results and the cardiac-CT results. In 184/324 (57%) women with negative ExECG, presence of CAD was diagnosed by means of cardiac-CT. Ninety out of 213 (42%) women with nondiagnostic ExECG did not show CAD. Furthermore, in 9/14 (64%) women with a positive ExECG, cardiac-CT showed no signs of CAD, and Table 3 shows that an additional 4/14 (29%) women only showed minimal stenosis on CCTA, whereas only one patient with a positive ExECG had a 70%–99% stenosis on CCTA. ExECG was negative in 26 (46%), inconclusive in 29 (52%), and positive in 1 (2%) of 56 women with >50% coronary artery stenosis on CCTA (see Fig. 1 for patient example). This is further detailed in Table 3, in which ExECG results are displayed against the grade of the most severe stenosis observed in each patient.

FIG. 1.

Example of a 50-year-old female evaluated for atypical chest pain and palpitations. Patient was known with hypercholesterolemia, smoking, and a strong family history of CAD. (a) Shows the exercise electrocardiogram (medians) that was normal in rest (R) and showed no STT changes during stress (S; max 100 W at peak stress, max heart rate 181/min, 106%) or during the recovery phase (Rec). No chest pain was reported by the patient during the exercise test. (b–d) Show a mixed plaque (large arrow) with a severe (70%–99%) lumen stenosis in the LCX in three directions, as captured in three planes (long axis, long axis, and short axis, respectively) with CCTA, performed to exclude CAD. (e) Shows the coronary angiogram confirming the lesion in the LCX, and (f) shows placement of a stent. (g) Displays the result directly after stenting of the affected coronary segment. CAD, coronary artery disease; R, rest; S, stress; Rec, recovery phase; LCX, left circumflex artery; LAD, left anterior descendens artery; IM, intermediate artery.

Table 2.

Exercise ECG Results Versus Coronary Artery Disease in Patients Subjected to CCTA and/or Calcium Scoring (n = 551)

	No CAD	CAD	Total
ExECG positive	9	5	14
ExECG negative	184^a	140	324
ExECG inconclusive	90	123	213
Total	283	268	551

One very young female with a calcium score of 0 was assumed to have no coronary artery stenosis and subsequent CCTA was omitted to minimize radiation burden.

ExECG, exercise ECG test; CAD, coronary artery disease; CCTA, coronary computed tomography angiography.

Table 3.

Exercise ECG Results Versus Maximal Stenosis Observed in Patients Subjected to Both Calcium Scoring and CCTA (n = 532)

	Maximal grade of stenosis observed in patients
	No stenosis	Minimal (0%–25%)	Mild (25%–50%)	Intermediate (50%–70%)	Severe (70%–99%)	Occlusion (100%)	Total
ExECG positive	9	4	0	0	1	0	14
ExECG negative	183	70	35	10	14	2	314
ExECG inconclusive	90	52	33	12	15	2	204
Total	282	126	68	22	30	4	532

Relationship of patient characteristics and CAD on cardiac-CT

In the entire cohort of 551 women, age, systolic blood pressure, and Duke Clinical Score showed a significant relationship with CAD by univariate regression analyses (Table 4). Since the univariate analyses returned a p < 0.20 for age, hypercholesterolemia, and systolic and diastolic blood pressure, these were further processed by multivariate analysis. In the multivariate regression analyses, only age proved to have a significant relationship with CAD in this cohort (p = 0.000).

Table 4.

Estimates of Correlation Between Patient Characteristics and Coronary Artery Disease on Cardiac-CT

	Univariate					Multivariate
	95% CI for Exp(B)					95% CI for Exp(B)
	B	Sig.	Exp(B)	Lower	Upper	B	Sig.	Exp(B)	Lower	Upper
Age	0.088	0.000	1.092	1.071	1.114	0.082	0.000	1.085	1.062	1.109
Family CAD	0.015	0.930	1.015	0.726	1.419	—	—	—	—	—
Smoking	−0.053	0.794	0.948	0.637	1.412	—	—	—	—	—
DM		0.436					—
DM1	−0.145	0.830	0.865	0.230	3.258		—
DM2	0.483	0.207	1.622	0.766	3.436		—
Hypercholesterolemia	0.344	0.069	1.411	0.973	2.046	0.061	0.768	1.063	0.709	1.592
RRsyst	0.024	0.000	1.025	1.014	1.035	0.009	0.226	1.009	0.995	1.023
RRdiast	0.012	0.144	1.012	0.996	1.027	−0.003	0.815	.997	0.976	1.019
BMI	0.001	0.936	1.001	0.968	1.036	—	—	—	—	—
Duke score	0.023	0.000	1.023	1.013	1.033

Data are ORs (odds ratios; 95% CI) or estimates of correlation.

Family CAD, presence of coronary artery disease in family history; DM, diabetes mellitus; DM1, diabetes mellitus type 1; DM2, diabetes mellitus type 2; RRsyst, systolic blood pressure before cardiac-CT; RRdiast, diastolic blood pressure before cardiac-CT; Duke score, Duke clinical score; BMI, body mass index.

Relationship of patient characteristics and ExECG

Table 5 shows the univariate and multivariate analyses of normal/nonnormal ExECG and patient characteristics. The characteristics age, exercise time, exercise load, maximal HR (maxHR), rest HR (restHR), systolic blood pressure at peak stress (X-RRsyst), diastolic blood pressure at peak stress (X-RRdiast), smoking, and BMI showed a significant correlation with ExECG test results (p < 0.05). Since the univariate analyses returned a p < 0.20 for age, exercise time, exercise load, maxHR, restHR, X-RRsyst, X-RRdiast, systolic blood pressure at rest (R-RRsyst), smoking, DM, and BMI, these were further processed by multivariate analysis. In the multivariate analyses, age, maxHR, exercise time, and R-RRsyst were significant (p = 0.000, p = 0.000, p = 0.038, and p = 0.011, respectively).

Table 5.

Estimates of Correlation Between Patient Characteristics and ExECG

	Univariate					Multivariate
	95% CI for Exp(B)					95% CI for Exp(B)
	B	Sig.	Exp(B)	Lower	Upper	B	Sig.	Exp(B)	Lower	Upper
Age	0.024	0.005	1.024	1.007	1.042	−0.059	0.000	0.943	0.914	0.972
maxHR	−0.080	0.000	0.923	0.910	0.936	−0.104	0.000	0.902	0.882	0.922
restHR	−0.038	0.000	0.963	0.949	0.976	−0.002	0.852	0.998	0.977	1.020
X-RRsyst	−0.016	0.000	0.984	0.978	0.991	0.010	0.118	1.010	0.997	1.023
X-RRdiast	−0.012	0.055	0.988	0.976	1.000	0.005	0.643	1.005	0.985	1.024
R-RRsyst	−0.005	0.177	0.995	0.987	1.002	−0.021	0.011	0.980	0.964	0.995
R-RRdiast	−0.009	0.217	0.991	0.976	1.005	—	—	—	—	—
Load	−0.020	0.000	0.980	0.975	0.985	0.004	0.392	1.005	0.994	1.015
ExECG time	−0.007	0.001	0.993	0.991	0.995	−0.003	0.038	0.997	0.994	1.000
Family CAD	−0.197	0.257	0.821	0.585	1.154	—	—	—	—	—
Smoking	0.506	0.013	1.659	1.113	2.476	−0.57	0.845	0.945	0.535	1.669
DM		0.188					0.780
DM1	1.098	0.127	2.971	0.735	12.013	0.457	0.587	1.580	0.303	8.237
DM2	0.396	0.293	1.485	0.711	3.105	−0.202	0.681	0.817	0.311	2.143
Hypercholesterolemia	0.012	0.951	1.012	0.695	1.473	—	—	—	—	—
BMI	0.044	0.013	1.045	1.009	1.083	−0.012	0.633	0.988	0.942	1.037
Duke score	0.000	0.995	1.000	0.991	1.009

Data are ORs (odds ratios; 95% CI) or estimates of correlation.

maxHR, maximal heart rate during ExECG test; restHR, heart rate before ExECG test; X-RRsyst, systolic blood pressure at peak stress; X-RRdiast, diastolic blood pressure at peak stress; R-RRsyst, systolic blood pressure before ExECG test; R-RRdiast, diastolic blood pressure before ExECG test; Load, maximal achieved exercise load; ExECG time, duration of the exercise; ECG, electrocardiogram.

Follow-up

The follow-up period of 56 patients with >50% stenosis on CCTA was 1381 ± 263 days (mean ± SD; range 945–1863 days). During follow-up, none of the women experienced myocardial infarction or cardiac death. 23/56 (41%) patients received revascularization by percutaneous coronary intervention (PCI) using stents only (of which 19 patients had a 70%–90% stenosis and 3 patients had a 50%–70% stenosis on CCTA), while 1/56 (2%) patients received CABG and 1/56 (2%) patients underwent both PCI with stenting and CABG.

Two out of 56 (4%) patients were diagnosed with significant CAD on CAG, but did not receive revascularization (1 refused and 1 was diagnosed with coronary artery dissection). A total of 7/56 (13%) patients showed nonsignificant lesions on CAG and the CAG of 3/56 (5%) patients with a positive CCTA was diagnosed as normal. No CAG was performed in 17/56 (30%) patients, and these patients were treated with medication only and were all stable during the follow-up period. Another 1/56 (2%) patients was not subjected to CAG since the encountered occlusion on CCTA was already known, and 1/56 (2%) patients was lost to follow-up.

The follow-up period of 324 female patients with negative ExECGs was 1460 ± 272 days (mean ± SD; range 897–1911 days). In this subgroup, no cardiac deaths were recorded during follow-up. However, 4/324 (1%) patients experienced myocardial infarction, of which two were revascularized. Revascularization by PCI with stenting was performed in 13/324 (4%) patients during the follow-up period.

Discussion

In this study, the suboptimal diagnostic power of ExECG in a patient cohort of women presenting with chest pain with a low-intermediate risk profile was confirmed. A substantial discrepancy between the ExECG test results and cardiac-CT diagnosis was observed.

CAD was detected using cardiac-CT in 184/324 (57%) women with negative ExECG test results, of which 26/184 (14%) showed >50% stenosis on cardiac-CT, which equals 8% of all women with negative ExECG results. In other words, in roughly 1 out of 12 women with a negative ExECG, significant CAD is missed assuming that the cardiac-CT can reliably act as a gold standard. Also, of the 14 women with a positive ExECG result, only one had significant CAD at cardiac-CT, whereas the majority had either no evidence of CAD or <50% stenosis on cardiac-CT (in 9/14 [64%] and 4/14 [29%] of the patients, respectively). Moreover, 128/213 (60%) women with nondiagnostic ExECG were diagnosed with CAD using cardiac-CT, of which 29/128 (23%) women had plaques with >50 stenosis on CCTA.

In this study, after exclusion of all patients with nondiagnostic ExECGs, and when considering ≥50% stenosis at cardiac-CT as the reference standard, the sensitivity, specificity, PPV, and NPV of ExECG are 3.7%, 95.7%, 7.1%, and 91.7%, respectively. When ≥70% stenosis at cardiac-CT is taken as the reference, it is 5.9%, 95.8%, 7.1%, and 94.9%, respectively.

The sensitivity and predictive value of ExECG appear suboptimal for proper diagnosis of CAD in women and is insufficient to guide (medical) treatment, although maximal achieved HR and exercise time proved to be independent predictors of CAD, besides other nonexercise-dependent variables such as age and systolic blood pressure in rest. It could be argued if cardiac-CT should be used as gold standard and alternatively, even more accurate diagnostic modalities (such as CAG) or extensive follow-up could be used.

In this cohort, revascularization was performed in 41% of the patients with >50% stenosis on CCTA. About 30% of the patients with less severe lesions did not proceed to CAG, were treated with medication, and proved to be stable during the follow-up period, during periodic visits at the outpatient clinic. Five percent of the patients diagnosed with >50% stenosis on CCTA proved to be normal at CAG, and can therefore be considered to have a false-positive CCTA. On the other hand, revascularization was performed by means of PCI with stenting in 4% of all women with negative ExECGs during the follow-up period.

There is a notable difference in the way women present and experience chest pain compared to men. The complaints are often described atypically,^14

–17 especially in young female patients, and are not always readily linked to CAD. In these patients, CAD may silently progress and eventually lead to major cardiac events,¹⁸ since this group is prone to be undertreated (less invasive or less aggressive) or receives no treatment at all.^19,20 Indeed, the overall decline of major cardiac events that is seen in the general population in the last couple of decades is not present in younger women,²¹ suggesting an underestimation of the condition in this particular subgroup.

While the average age of the first myocardial infarction in female patients is higher than in men, the outcome is worse in women, that is, mortality is higher after myocardial infarction than in men.¹⁸ Obviously, reliable methods to diagnose or rule out CAD in female patients in an early stage are necessary to initiate preventive medication timely and prevent major cardiac adverse events from occurring.

The value of ExECG in women has been under debate for many years and nondiagnostic or even erroneous test results are frequently encountered. Although combining ExECG test results with other parameters such as exercise capacity, percentage of age-predicted exercise capacity, chronotropic response, HR recovery, blood pressure response, and the Duke Treadmill Score⁵ can all be used to increase the reliability of the test,⁴ there is an urgent need for better CAD assessment in women, especially in symptomatic female patients with a low-to-intermediate risk for significant CAD.

In a report²² published by Patel et al. in 2014, it was shown, from a very large amount of data retrieved from the USA national NCDR's CathPCI Registry, that about two-thirds of the patients had undergone noninvasive tests, including ExECG, before elective CAG. That study also mentioned that in nearly one in six patients with nonobstructive CAD (<50% stenosis) at CAG, intermediate or high-risk findings were detected with noninvasive tests. Also, more than 57% of the patients with low-risk (but positive) findings on noninvasive tests were diagnosed with nonobstructive CAD by means of CAG, whereas nearly one-third of patients with high-risk findings on noninvasive tests had nonobstructive disease.

In this study, all except one patient with positive ExECGs had nonobstructive CAD on CCTA (nine patients with no and four patients with minimal stenosis; Table 3). Moreover, in this study, only 1/56 patients with >50% stenosis on CCTA was documented with a positive ExECG. In 29/56 (52%) patients with >50% stenosis on CCTA, the ExECG was inconclusive and ExECG was negative in the remainder of the patients.

While in the study of Patel, various types of noninvasive testing were evaluated in a study group containing both sexes, with undefined pretest likelihood for significant coronary disease, the substantially worse performance of ExECG in the cohort of this study compared to CCTA results underlines the notion that the value of ExECG in women with low-to-intermediate risk for significant CAD is at least questionable.

Preventive therapeutics for patients with stable CAD include low-dose aspirin medication, which is a class IA recommendation in male and female patients with stable CAD.^2,23 Due to the lack of reliability of ExECG in women, initiation or termination of low-dose aspirin should not be based on ExECG test results. However, suboptimal postcardiac-CT aspirin use has been described in up to 30% of the patients with CAD in several studies for a variety of reasons.^24

–27 It has not been studied whether overestimation of CAD by cardiac-CT leads to inappropriate low-dose aspirin use.

Similarly, preventive medication such as statin treatment should not be based solely on the results of ExECG tests. While statin treatment reduces CAD morbidity and mortality in patients at an increased risk of cardiac events,^28
–30 a considerable general undertreatment with statins was demonstrated in a recent study by Nilsson et al., especially among women and the elderly.³¹ In this study, baseline low-dose aspirin treatment, baseline statin treatment, and treatment with a combination of the two medications were reported in 225/551 (41%), 179/551 (33%), and 144/551 (26%) patients, respectively. Cardiac-CT results led to low-dose aspirin initiation in 46/551 (8%) women, whereas statins were started in 103/551 (19%) women. Furthermore, low-dose aspirin was discontinued in 101/551 (18%) women, while statins were stopped in 65/551 (12%) women.

In this study, the mean radiation dose of the cardiac-CT was 2.1 ± 1.9 (range 0.31–16.3), which was substantially lower than the cardiac-CT dose of 8–9 mSv in the ROMICAT-2 trial.³² In women, the radiation dose to, in particular, the breasts should be kept as low as possible. Breast doses of up to 0.01 Gy delivered to patients <35 years may increase the lifetime risk of breast cancer by 13.6%.^33
–35 Using dose-saving algorithms and state-of-the-art CT technology, typical radiation doses of less than 3 mSv can be achieved.^8,9

As in most publications, we used a factor of 0.014 mSv/(mGy·cm) for DLP to effective dose (mSv) conversion. Based on revised tissue weighting factors in the International Commission on Radiological Protection (ICRP) publication 103 (2007), higher conversion factors up to 0.028 mSv/(mGy·cm) have also been suggested.^36,37 Taking this into account, usage of optimal scanning and dose-reduction techniques is of utmost importance to minimize the radiation delivered to breasts.

This study has several limitations. The study is a single-center study reflecting regular clinical practice and a referral bias should therefore be taken into account, since only women who had undergone both ExECG and cardiac-CT were included in the study. Women with normal ExECGs and without further suspicion of having significant CAD may not have pursued CCTA, and the pretest probability of CAD could therefore be higher in the investigated cohort than in all women (including those that did not receive CCTA). Likewise, women with convincing positive ExECGs may also have proceeded to other types of noninvasive testing (such as myocardial perfusion PET) or to CAG directly.

Also, as outlined above, a definite gold standard is missing. CCTA is known to overestimate the degree of stenosis, especially in calcified coronary plaques. Hence, the number of patients with >50% coronary artery stenosis could be overestimated in this study and remains uncertain, as no CAG was performed in 30% of these patients. The low fraction of patients with positive ExECGs is another limitation. Furthermore, this study did not focus on long-term outcome in women with and without CAD, although follow-up data were obtained for two groups of interest. Despite these limitations, since we performed the analysis in an all-comers patient cohort, this study is an actual reflection of daily clinical practice.

In conclusion, ExECG failed to detect CAD in more than half of low-intermediate risk women presenting with chest pain and in almost half of the patients with >50% stenosis at CCTA. Importantly, no CAD was detected by cardiac-CT in 64% of women with a positive ExECG. ExECG is therefore questionable as a diagnostic strategy in women with low-to-intermediate risk to detect CAD. The low radiation exposure (2.1 mSv) signifies the probability of cardiac-CT to become a suitable alternative in this subset of female patients. Further studies are mandated to explore the future diagnostic usability of this imaging technique and long-term outcome in women.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Wenger

. Juggling multiple guidelines: A woman's heart in the balance. J Women's Health, 2016; 25:213–221.

Montalescot

, Sechtem

, Andreotti

, et al. ESC guidelines on the management of stable coronary artery disease: The Task Force on the management of stable coronary artery disease of the European Society of Cardiology. Eur Heart J, 2013; 34:2949–3003.

Morise

, Diamond

. Comparison of the sensitivity and specificity of exercise electrocardiography in biased and unbiased populations of men and women. Am Heart J, 1995; 130:741–747.

Kohli

, Gulati

. Exercise stress testing in women: Going back to the basics. Circulation, 2010; 122:2570–2580.

Mark

, Hlatky

, Harrell

, Lee

, Califf

, Pryor

. Exercise treadmill score for predicting prognosis in coronary artery disease. Ann Intern Med, 1987; 106:793–800.

Hadamitzky

, Täubert

, Deseive

, et al. Prognostic value of coronary computed tomography angiography during 5 years of follow-up in patients with suspected coronary artery disease. Eur Heart J, 2013; 34:3277–3285.

Min

, Dunning

, Lin

, et al. Age- and sex-related differences in all-cause mortality risk based on coronary computed tomography angiography findings results from the International Multicenter CONFIRM (Coronary CT Angiography Evaluation for Clinical Outcomes: An International Multicenter Registry) of 23,854 patients without known coronary artery disease. J Am Coll Cardiol, 2011; 58:849–860.

Achenbach

, Marwan

, Ropers

, et al. Coronary computed tomography angiography with a consistent dose below 1 mSv using prospectively electrocardiogram-triggered high-pitch spiral acquisition. Eur Heart J, 2010; 31:340–346.

Ghoshhajra

, Engel

, Major

, et al. Evolution of coronary computed tomography radiation dose reduction at a tertiary referral center. Am J Med, 2012; 125:764–772.

10.

Agatston

, Janowitz

, Hildner

, Zusmer

, Viamonte

, Detrano

. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol, 1990; 15:827–832.

11.

Bogaard

, van der Zant

, Knol

RJJ

, et al. High-pitch prospective ECG-triggered helical coronary computed tomography angiography in clinical practice: Image quality and radiation dose. Int J Cardiovasc Imaging, 2014; 31:1–9.

12.

van der Zant

, Wondergem

, Lazarenko

, et al. Ruling out coronary artery disease in women with atypical chest pain: Results of calcium score combined with coronary computed tomography angiography and associated radiation exposure. J Women's Health, 2015; 24:550–556.

13.

Bayliss

. Duke Clinical Score: Prediction of coronary heart disease in a patient with chest pain, 1999. www.zunis.org/Duke Chest Pain - CAD Predictor.htm Accessed March 21 2017 .

14.

Granot

, Goldstein-Ferber

, Azzam

. Gender differences in the perception of chest pain. J Pain Symptom Manage, 2004; 27:149–155.

15.

Steingart

, Packer

, Hamm

, et al. Sex differences in the management of coronary artery disease. Survival and Ventricular Enlargement Investigators. N Engl J Med, 1991; 325:226–230.

16.

Ghadri

, Sarcon

, Jaguszewski

, et al. Gender disparities in acute coronary syndrome: A closing gap in the short-term outcome. J Cardiovasc Med, 2015; 16:355–362.

17.

Pagidipati

, Peterson

. Acute coronary syndromes in women and men. Nat Rev Cardiol, 2016; 13:471–480.

18.

Dittrich

, Gilpin

, Nicod

, Cali

, Henning

, Ross

. Acute myocardial infarction in women: Influence of gender on mortality and prognostic variables. Am J Cardiol, 1988; 62:1–7.

19.

Canto

, Rogers

, Goldberg

, et al. Association of age and sex with myocardial infarction symptom presentation and in-hospital mortality. JAMA, 2012; 307:813–822.

20.

Davis

, Diamond

, Montgomery

, Krishnan

, Eagle

, Jackson

. Acute coronary syndrome in young women under 55 years of age: Clinical characteristics, treatment, and outcomes. Clin Res Cardiol, 2015; 104:648–655.

21.

Kawamoto

, Davis

, Duvernoy

. Acute coronary syndromes: Differences in men and women. Curr Atheroscler Rep, 2016; 18:73.

22.

Patel

, Dai

, Hernandez

, et al. Prevalence and predictors of nonobstructive coronary artery disease identified with coronary angiography in contemporary clinical practice. Am Heart J, 2014; 167:846–852.

23.

Fihn

, Gardin

, Abrams

, et al. ACCF/AHA/ACP/AATS/PCNA/SCAI/STS guideline for the diagnosis and management of patients with stable ischemic heart disease. J Am Coll Cardiol, 2012; 60:e44–e164.

24.

Cheezum

, Hulten

, Smith

, et al. Changes in preventive medical therapies and CV risk factors after CT angiography. JCMG, 2013; 6:574–581.

25.

Ovrehus

, Bøtker

, Jensen

, Munkholm

, Johnsen

, Nørgaard

. Influence of coronary computed tomographic angiography on patient treatment and prognosis in patients with suspected stable angina pectoris. Am J Cardiol, 2011; 107:1473–1479.

26.

Bom

, van der Zee

, Cornel

, van der Zant

, Knol

RJJ

. Diagnostic and therapeutic usefulness of coronary computed tomography angiography in out-clinic patients referred for chest pain. Am J Cardiol, 2015; 116:1–7.

27.

Hachamovitch

, Nutter

, Hlatky

, et al. Patient management after noninvasive cardiac imaging. J Am Coll Cardiol, 2012; 59:462–474.

28.

Baigent

, Blackwell

, Emberson

, et al. Efficacy and safety of more intensive lowering of LDL cholesterol: A meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet, 2010; 376:1670–1681.

29.

Baigent

, Keech

, Kearney

, et al. Efficacy and safety of cholesterol-lowering treatment: Prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet, 2005; 366:1267–1278.

30.

Mihaylova

, Emberson

, Blackwell

, et al. The effects of lowering LDL cholesterol with statin therapy in people at low risk of vascular disease: Meta-analysis of individual data from 27 randomised trials. Lancet, 2012; 380:581–590.

31.

Nilsson

, Samuelsson

, Söderström

, Mooe

. Low use of statins for secondary prevention in primary care: A survey in a northern Swedish population. BMC Fam Pract, 2016; 17:110.

32.

Truong

, Hayden

, Woodard

, et al. Sex differences in the effectiveness of early coronary computed tomographic angiography compared with standard emergency department evaluation for acute chest pain: The rule-out myocardial infarction with Computer-Assisted Tomography (ROMICAT)-II Trial. Circulation, 2013; 127:2494–2502.

33.

Hopper

, King

, Lobell

, TenHave

, Weaver

. The breast: In-plane x-ray protection during diagnostic thoracic CT—shielding with bismuth radioprotective garments. Radiology, 1997; 205:853–858.

34.

Liu

, Goodsitt

, Chan

. Normalized average glandular dose in magnification mammography. Radiology, 1995; 197:27–32.

35.

McCollough

, Liu

. Breast dose during electron-beam CT: Measurement with film dosimetry. Radiology, 1995; 196:153–157.

36.

Christner

, Kofler

, McCollough

. Estimating effective dose for CT using dose-length product compared with using organ doses: Consequences of adopting International Commission on Radiological Protection publication 103 or dual-energy scanning. AJR Am J Roentgenol, 2010; 194:881–889.

37.

Paul

, Jacobi

, Bazrafshan

, Farshid

, Vogl

. Effect of contrast material on radiation dose in an adult cardiac dual-energy CT using retrospective ECG-gating. Health Phys, 2013; 105:156–164.