Article, Cardiology

Diagnostic yield of coronary angiography in patients with acute chest pain: role of noninvasive test

a b s t r a c t

Objectives: This study investigated the diagnostic yield of invasive Coronary angiography and the impact of Noninvasive test (NIV) in patients presented to emergency department (ED) with acute chest pain. Methods: Patients 50 years or older who visited ED with acute chest pain and underwent CAG were identified retrospectively. Those with ischemic electrocardiogram, elevated cardiac enzyme, known coronary artery disease (CAD), history of cardiac surgery, renal failure, or allergy to radiocontrast were excluded. Diagnostic yields of CAG to detect significant CAD or differentiate the need for revascularization were analyzed according to whether NIV was performed and its result.

Results: Among the total 375 consecutive patients, significant CAD was observed in 244 (65.1%). Diagnostic yields of CAG were higher in patients who underwent NIV before CAG, but the discriminative effect was modest (59.7% vs 70.7% [P = .026] for detection of CAD; 45.0% vs 50.5% [P = .285] for revascularization). Positive results of NIV were significantly associated with the presence of CAD and the need for revascularization, when compared with patients without NIV or patients with negative results (P b .001, respectively).

Conclusion: The diagnostic yield of CAG was only 65% in low- to intermediate-risk ED patients with acute chest pain. Performing of NIV provided only modest improvement in diagnostic yield of CAG. The unexpectedly low diagnostic yield might be attributable to the underuse of NIV and misinterpretation of physicians. We suggest the use of NIV as a gatekeeper to discriminate patients who require CAG and/or revascularization, and for this, better risk stratification and appropriate application of NIV are required.

Introduction

Acute chest pain is one of the most common chief concerns of patients who present to the emergency department (ED) [1]. Because appropriate triage for patients with acute chest pain is important, physicians must estimate the pretest probability by weighing the symptoms, risk factors, level of cardiac enzymes, and the results of noninvasive tests (NIVs) [2]. Traditionally, cardiac stress tests such as exercise treadmill test (TMT) and myocardial single-photon emission computed tomography (SPECT) are recommended in low- to intermediate-risk patients [2-4]. In contrast, coronary angiography (CAG) should be performed without NIV in high-risk patients, not to delay the immediate diagnosis and treatment [5]. Beside the use of

? Source of funding: This study was supported by the National Research Foundation of Korea grant funded by the Ministry of Education, Science and Technology (MEST) (A070001).

?? Conflicts of interest: There is no conflict of interest regarding this study.

* Corresponding author. Department of Internal Medicine, Seoul National University College of Medicine, Cardiovascular Center, Seoul National University Hospital, 101 Daehak- ro, Jongno-gu, Seoul 110-744, South Korea. Tel.: +82 2 2072 1963; fax: +82 2 2072 2577.

E-mail address: [email protected] (Y.-J. Kim).

stress tests, coronary CT angiography (CCTA) is especially useful when cardiac enzymes and Electrocardiograms have not shown any abnormalities [6]. Predictive value for long-term prognosis and cost- effectiveness of CCTA has also been confirmed by a series of randomized trials and meta-analyses [7-12].

Because the risk of complications and the resultant costs are inescapable part of CAG, we should improve the diagnostic yield of CAG, so as to perform this invasive procedure in those with necessity. Although there have been tremendous progress in the use of NIV, the diagnostic yield of invasive CAG is unsatisfactory. In a large-scale clinical study, a surprisingly low diagnostic yield of CAG was revealed, which was only slightly more than one-third of the patients [13]. Moreover, the diagnostic yield of CAG showed considerable hospital variability [14]. Thus, better strategies are needed to minimize unnecessary invasive procedures and the risk of complications.

Noninvasive tests including stress tests and CCTA have the potential to reduce the unwarranted use of CAG [15,16]. However, there have not been many studies investigated the diagnostic yield of CAG in patients who visited ED with acute chest pain, according to the performing of NIV and their results. We investigated the diagnostic yields of CAG for the detection of significant coronary artery disease

(CAD) and differentiation of the need for revascularization, as well as the impact of NIV on the diagnostic yields, among low- to intermediate-risk ED patients with acute chest pain.

Methods

Study design

This was an observational single-center study of patients with acute chest pain who underwent CAG to diagnose or rule out significant CAD. Among patients who visited Seoul National Univer- sity Hospital ED with acute chest pain from January 2005 to December 2011, we identified 375 consecutive patients 50 years or older who were admitted for CAG without elevated cardiac enzyme, ST-segment depression on ECG, or cardiogenic shock at presentation. Patients were also excluded if they had established CAD or history of cardiac surgery. Baseline demographic variables were identified based on the medical records. Traditional risk factors for CAD and Thrombolysis in Myocardial Infarction risk scores were assessed according to the current guidelines [2,17]. This study was approved by the institutional review board of Seoul National University Hospital (H-1108-108-375).

Coronary angiography

Coronary angiography was performed by experienced cardiolo- gists in our hospital, using Judkins technique via a transfemoral or transradial approach. At least 2 projections were identified to evaluate a right coronary artery lesion, and at least 6 projections were identified to evaluate a left coronary artery lesion. Stenosis severity was determined by quantitative coronary analysis (version 3.3; Philips, Einthoven, the Netherlands), which indicates the maximal luminal diameter stenosis on a projection. Significant lesion was defined as more than 50% stenosis of major epicardial coronary arteries.

Coronary CT angiography

Patients in this study underwent 64-slice multidetector CT (SOMATOM Sensation 64 and SOMATOM Definition; Siemens Medical Solutions, Forchheim, Germany). Patients with a prescanning heart rate of 65 beats/min or higher were given 50 to 100 mg of oral metoprolol 45 to 60 minutes before the CT examination, unless the subject had a contraindication to ?-blockers. During the image acquisition, 60 to 80 mL of contrast was injected, followed by a saline flush. Coronary CT angiography was performed using the retrospec- tive ECG-gated mode with ECG pulsing, and the scans were usually performed from the diaphragm to the level of the tracheal bifurcation in the caudocranial direction. All CCTA results were transferred to an external 3-dimensional workstation and analyzed independently by 2 experienced radiologists. Positive result on CCTA was defined as any more than 50% stenosis of the left main coronary artery or other major epicardial coronary arteries.

Stress tests

Stress tests including TMT and myocardial SPECT were performed

formed 90 minutes after stress. Short- and long-axis images were reconstructed, and the image interpretation was performed by 2 experienced nuclear cardiologists based on the semiquantitative polar maps of perfusion. Perfusion defects from each segment were assessed on the stress image (segmental tracer activity, b 75% of maximum) and divided into reversible (>=10% increase in tracer uptake on the resting image) or fixed perfusion defects. Positive result on myocardial SPECT was defined as reversible or fixed perfusion defects.

Statistical analysis

Results were presented as means +- SD or frequencies with percentages. Group comparisons were performed using the Student t test or crosstabs. The ?² or the Fisher exact test was used for categorical variables. Diagnostic yields of CAG for detection of significant CAD and revascularization treatment were analyzed using crosstabs. Diagnostic yields were also analyzed according to the types of NIV and their results, which was performed as an initial test to the patients on ED presentation. Positive results of NIV were determined as any of CCTA, myocardial SPECT, or TMT showing signs of significant CAD. All statistical analyses were performed with software SPSS 20.0 (IBM Corp, Armonk, NY), and a P value less than

.05 was considered statistically significant.

Results

Baseline characteristics

Among 375 patients, 184 (49.1%) underwent NIV before CAG (Table 1). Prevalence of dyslipidemia was more frequent in patients without NIV (36.6% vs 23.4%, P = .005). Other baseline characteristics were not different between the groups. The traditional risk factors and TIMI risk score were not significantly different between the 2 groups.

Diagnostic yield of CAG and impact of NIVs

Results of CAG were analyzed according to the performing of NIV (Table 2). In total, significant CAD was confirmed in 244 (65.1%) patients, and revascularization treatments including percutaneous coronary intervention (PCI) and coronary artery bypass graft (CABG) were performed in 179 (47.7%) patients. Diagnostic yields of CAG for detection of significant CAD were higher in patients with NIV (59.7% vs 70.7%, P = .026). Revascularization treatments were more frequently performed in those who underwent NIV before CAG, without statistical significance (45.0% vs 50.5%, P = .285). Compared with those without NIV or with negative results on NIV, patients with positive results on NIV demonstrated a higher diagnostic yield of CAG for detection of CAD (85.9% vs 59.7% of those without NIV [P b .001], vs 19.0% of those with negative results on NIV [P b .001]; Fig. 1).

Table 1

Baseline characteristics

Without NIV (n = 191) With NIV (n = 184) P

Age (y) 67.0 +- 9.0 66.7 +- 8.3 .815

Sex (male) 88 (46.1%) 93 (50.5%) .386

Comorbidity

upon the physician’s discretion. Treadmill test was performed	Diabetes mellitus	47 (24.6%)	45 (24.5%)	.973
following the Bruce protocol, and the results were interpreted	Hypertension	123 (64.4%)	119 (64.7%)	.955
according to the American Heart Association guideline [18]. Single-	Dyslipidemia	70 (36.6%)	43 (23.4%)	.005
photon emission computed tomography was performed with phar- macologic stress, using a 1-day ECG-gated stress-rest protocol. Technetium-99m was injected intravenously, and a resting image	CVA Current smoker Family history of CAD BMI (kg/m²)	15 (7.9%) 49 (25.7%) 19 (9.9%) 24.0 +- 3.4	13 (7.1%) 34 (18.5%) 24 (13.0%) 24.5 +- 2.9	.772 .094 .347 .100
was acquired. Pharmacologic stress was induced by infusion of	CAD risk factor	1.61 +- 0.98	1.44 +- 0.96	.086
adenosine (0.14 mg kg^-1 min^-1). Technetium-99 m was injected 3	TIMI risk score	1.43 +- 1.01	1.33 +- 0.85	.285
minutes after stress, and gated technetium-99m SPECT was per- Abbreviations: BMI, body mass index; CVA, cerebrovascular accident.

Table 2 Results of CAG
	Without NIV	With NIV	P
	(n = 191)	(n = 184)
Normal or insignificant	77 (40.3%)	54 (29.3%)	.003
CAD without revascularization	28 (14.7%)	37 (20.1%)
PCI	80 (41.9%)	72 (39.1%)
CABG	6 (3.1%)	21 (11.4%)
1VD	60 (31.4%)	55 (29.9%)	.001
2VD	36 (18.8%)	30 (16.3%)
3VD	18 (9.4%)	45 (24.5%)
LM disease	8 (4.2%)	12 (6.5%)	.315
Detection of CAD^a	114 (59.7%)	130 (70.7%)	.026
Revascularization^b	86 (45.0%)	93 (50.5%)	.285

Abbreviations: LM, left main; VD, vessel disease.

^a Significant CAD by CAG was defined as more than 50% stenosis of left main coronary artery or other epicardial coronary arteries.

^b Revascularization treatment including PCI or CABG.

Diagnostic yield for revascularization treatment showed similar results (63.4% vs 45.0% of those without NIV [P b .001], vs 7.1% of those without negative results on NIV [P b .001]).

We analyzed the diagnostic yields of CAG according to the types of NIV (Fig. 2 and Table 3). Diagnostic yields for detection of significant CAD were 70.7% in patients with NIV, 71.4% in patients with CCTA, 71.1% in patients with myocardial SPECT, and 66.7% in patients with TMT. Diagnostic yields for revascularization were 50.5%, 51.8%, 44.4%, and 55.6%, respectively. Although the patients who underwent each NIV demonstrated higher diagnostic yields than did those without NIV, the improvement in discrimination was modest.

Diagnostic yields of CAG were also compared according to the results of each NIV (Fig. 3 and Table 3). Among those who underwent CCTA as an initial NIV, positive results were observed in 87 patients. Coronary angiography revealed significant CAD in 89.7% among those with positive results on CCTA and 8.0% among negative results. Revascularization treatments were performed in 66.7% and 0.0%, respectively. Positive results of myocardial SPECT were observed in 34 patients. Significant CAD by CAG was confirmed in 85.3% among those with positive results on myocardial SPECT and 27.3% in those with negative results. Among the patients with positive results on myocardial SPECT, revascularization treatments were performed in 55.9%, whereas it was 9.1% among those with negative results. Positive results of TMT yielded 71.4% of significant CAD by CAG and 61.9% of revascularization treatment, whereas negative results yielded 50.0% and 33.3%, respectively.

Fig. 1. Impact of NIVs on diagnostic yields of CAG. Diagnostic yields of CAG for detection of significant CAD and differentiation of the need for revascularization treatment were shown, according to the performing of NIV. Diagnostic yield for detection of CAD was higher in patients who underwent NIV before CAG, showing about 10% of improvement (left, white). Regarding revascularization treatment, the NIV group showed a slightly higher diagnostic yield (left, black). Patients with positive results on NIV showed a significantly higher diagnostic yield compared with patients with negative results or those without NIV (right, white and black). *P b .05 vs those without NIV. ⁺P b .001 vs those without NIV. ^?P b .001 vs those with negative result on NIV.

Fig. 2. Diagnostic yields of CAG according to the types of NIVs. Diagnostic yields of CAG for detection of significant CAD (white) and revascularization treatment (black) of each NIV were analyzed, in comparison with those without NIV. Although the diagnostic yields for detection of CAD were higher in those who underwent NIV, the improvements in discrimination were only modest.

Discussion

In this study, we investigated the diagnostic yield of CAG among 375 consecutive patients who presented to ED with acute chest pain without definite abnormalities in ECG and cardiac enzyme level. We found that diagnostic yield for detection of significant CAD was only 65.1% in all patients. Although the yield was higher in patients who had undergone NIV before CAG, the efficacy was only modest, resulting in about 10% improvement in the diagnostic yields. This is a disappointing result considering the recent development of NIV. Diagnostic yields should not be 100% because of the risks of missed diagnosis; however, diagnostic yields of 65.1% imply a lot of room for improvement in our real-world practice.

Diagnostic tests for patients with acute chest pain

Acute chest pain is a frequent complaint in the ED and also is an important symptom that requires expeditious and meticulous diagnostic tests and Clinical decisions [1,2]. Delayed or missed diagnosis causes a substantial increase in morbidity and mortality, especially in the case of acute coronary syndrome [5]. In patients with unstable angina/Non-ST-segment elevation myocardial infarction, assessment of pretest probability and risk stratification are manda- tory, and NIV is useful when cardiac biomarkers and resting ECG results are negative [2]. In these low- to intermediate-risk patients, stress tests are recommended and CCTA is also appropriate as an alternative to stress tests, based on their accuracy, cost-effectiveness, and prognostic information [2-4].

Table 3 Diagnostic yields of CAG by NIVs
	Detection of significant CAD^a	Revascularization treatment^b
Without NIV (n = 191)	114 (59.7%)	86 (45.0%)
NIV (n = 184)	130 (70.7%)	93 (50.5%)
Positive result (n = 142)	122 (85.9%)	90 (63.4%)
Negative result (n = 42)	8 (19.0%)	3 (7.1%)
CCTA (n = 112)	80 (71.4%)	58 (51.8%)
Positive result (n = 87)	78 (89.7%)	58 (66.7%)
Negative result (n = 25)	2 (8.0%)	0 (0.0%)
Myocardial SPECT (n = 45)	32 (71.1%)	20 (44.4%)
Positive result (n = 34)	29 (85.3%)	19 (55.9%)
Negative result (n = 11)	3 (27.3%)	1 (9.1%)
TMT (n = 27)	18 (66.7%)	15 (55.6%)
Positive result (n = 21)	15 (71.4%)	13 (61.9%)
Negative result (n = 6)	3 (50.0%)	2 (33.3%)

^a Significant CAD by CAG was defined as more than 50% stenosis of left main coronary artery or other epicardial coronary arteries.

^b Revascularization treatment including PCI or CABG.

Fig. 3. Diagnostic yields of CAG according to the results of NIVs. In patients with positive results or negative results of each NIV, diagnostic yields of CAG for detection of CAD (white) and revascularization treatment (black) were demonstrated.

Cardiac stress tests including TMT and myocardial SPECT have been the diagnostic test of choice in low- to intermediate-risk patients. Diagnostic accuracy of TMT was provided in a study, reporting 99.4% of specificity and 98.7% of negative predictive value (NPV) [19]. In addition, immediate TMT in low-risk patients with acute chest pain showed acceptable safety [20]. In a randomized controlled trial by Udelson et al [16], myocardial SPECT reduced unnecessary hospitalizations without reducing appropriate admission in patients with Acute cardiac ischemia and nondiagnostic ECG. Moreover, myocardial SPECT is cost-effective and also provides prognostic information among patients with acute chest pain [21].

Recently, many studies have been conducted to investigate the diagnostic performance of CCTA in patients with acute chest pain. The Rule Out Myocardial Infarction using Computer Assisted Tomography trial demonstrated that 64-slice CCTA has a sensitivity of 60% to 100%, a specificity of 50% to 90%, and an NPV of 98% to 100% [8]. The 2-year outcome of the Rule Out Myocardial Infarction using Computer Assisted Tomography trial showed the excellent prognos- tic value of the CCTA results [9]. For patients with possible acute coronary syndrome, CCTA-based strategy improved the efficiency of decision making and allowed safe discharge from ED [22,23]. Because CCTA achieved recognition in recent years, multiple guidelines recommend CCTA for patients with low-to-intermediate pretest probability of CAD or in the absence of cardiac enzyme elevation and ischemic ST changes on ECG when the patient presented with acute symptoms [2,3,6].

Diagnostic yield of CAG and the impact of NIVs

Despite the excellent diagnostic performance of NIV and well- established guidelines on acute chest pain, considerable real-practice problems remain unsolved. In particular, the diagnostic yield of elective CAG is low according to a study by Patel et al [13], in which the diagnostic yield was less than 40%, and NIV including ECG, stress tests, echocardiography, CCTA, and other heart scans contributed only a modest effect on the ability to predict the presence of CAD. This important study reminds us that “inappropriate” invasive procedures are frequent, which lead to the avoidable risk of complications. Another large-scale study of 565 504 patients by Douglas et al [14], the diagnostic yield of CAG varied considerably among hospitals. Up to one-third of CAG procedures might not be required if the strategies for patient selection, treatment, and testing patterns were improved. Although the above 2 studies excluded patients with acute coronary syndrome, their clinical implications for appropriate use of CAG are

quite clear. It should not be confined to the patients undergoing elective diagnostic CAG [13,14], but also applied to the patients with acute chest pain for whom CAG is considered. As reported in a retrospective study by Hamid et al [24], a significant proportion of ED patients with acute chest pain are suitable for CCTA. These studies strongly imply that better Risk stratification strategies and appropriate applications of NIV are needed in patients with acute chest pain. Our study focused on the diagnostic yield of CAG and the role of NIV among low- to intermediate-risk patients who presented to the ED with acute chest pain.

In the present study, diagnostic yields of CAG for detection of significant CAD were higher in patients who underwent NIV before CAG (Fig. 1). Revascularization treatments were slightly more frequent in the NIV group. Considering the excellent diagnostic performance of NIV [8,16,25], diagnostic yields of CAG should be significantly higher in patients who underwent NIV than those not; however, the discriminative effect of NIV was only modest. We also analyzed the diagnostic yields according to the types of NIV and their results (Figs. 2 and 3). Patients who underwent each NIV showed slightly higher diagnostic yields than did those without NIV, and those with positive results on CCTA or myocardial SPECT had significant higher diagnostic yields, whereas TMT did not show a significant result because of their small size. However, the important thing is that the increments in diagnostic yields were only about 10%, even when analyzed separately by the types of NIV.

These results imply that NIV is not applied for appropriate patients in real-world practice. As recommended in the current guidelines, NIV is helpful and appropriate in patients with low-to-intermediate pretest probability of CAD and in patients with normal ECG and cardiac enzymes [2,3,6]. When reviewed retrospectively, most of study populations were appropriate candidates for NIV because we included patients who presented to ED with acute chest pain and excluded those with elevated cardiac enzymes or ischemic ST-segment changes. However, only 50% of patients underwent NIV before CAG. Moreover, it seemed that the results of NIV were not appropriately reflected into the real-world practice. Substantial patients underwent unnecessary CAG despite the negative results on NIV.

Baseline demographic factors were not significantly different between the 2 groups, except for dyslipidemia, which was more frequent in patients without NIV. Traditional risk factors and TIMI risk score were not different, whereas they were tended to be higher in patients without NIV. Although these slight differences could have affected the physician’s preference on NIV, still the diagnostic yields in the NIV group are unsatisfactory.

Noninvasive tests to improve the diagnostic yield of CAG

As reported previously, NPV of CCTA is near 100% [7,8], and thus, negative results of CCTA warrant safe discharge from ED [23]. Although not as excellent as CCTA, myocardial SPECT and TMT have acceptable and comparable diagnostic performance [19,26,27]. Thus, we could assume that low- to intermediate-risk patients with acute chest pain in this study would show significantly higher diagnostic yields of CAG when they underwent NIV before CAG. Based on the results of this study and also the previous reports, we suggest that more patients should have performed NIV in ED as a gatekeeper test. In a study by Moscariello et al [15], both CCTA and CAG were performed in symptomatic patients who showed positive results on SPECT. Because the clinical decisions from CCTA and CAG regarding the need for revascularization and the selection of a revascularization strategy were not different, the authors suggested that CCTA would have the potential to minimize the unnecessary CAG. However, they could not provide the difference in diagnostic yield of CAG according to whether CCTA was performed and the result of CCTA because of the design of study. In contrast, our study revealed that diagnostic yields of CAG were only slightly higher when NIVs were performed, and it could be improved by the appropriate use of NIV. We believe that our results provide another evidence of the potential role of NIV, which discriminate patients who require invasive procedures for further treatment from those who could be discharged safely. In addition, our study included those who visited ED with acute chest pain, reflecting the real-world practice. Because the impact of NIV on the diagnostic yield of CAG in these patients was not

studied before, our study could suggest more practical evidence.

Study limitations

This study is subject to following limitations. First, it was a retrospective single-center study. All patients underwent NIV or CAG based on the discretion of attending physicians without previously standardized protocol, and clinical reasoning for each case could not be counted. However, the diagnostic yields of CAG and the impact of NIV should be interpreted on the aspect of real-world practice, regarding the clinical reasoning for each case as a part of overall results. Because we included the consecutive ED patients with acute chest pain who admitted for CAG, the conclusion of our study might not be affected or biased because of the design of this study. Instead, we believe that our study would provide important evidence for real-world practice. Moreover, there was no significant difference in overall baseline characteristics between the 2 groups: patients without NIV vs with NIV. This feature enhances the value of our study, emphasizing the impact of NIV on diagnostic yields of CAG. Second, patients who were discharged from ED without performing CAG were not included in this study. Although we could not provide the exact values of diagnostic performance of NIV, the main focus of this study was the diagnostic yield of CAG, which might give valuable insights.

Conclusion

The diagnostic yield of CAG was only 65% in low- to intermediate- risk ED patients with acute chest pain. Performing of NIV provided only modest improvement in the diagnostic yield of CAG. The unexpectedly low diagnostic yield of CAG might be attributable to the underuse of NIV and misunderstandings of physicians. We suggest the use of NIV as a gatekeeper to discriminate patients who require CAG and/or revascularization, and for this, better risk stratification and appropriate application of NIV are required.

Acknowledgments

The authors gratefully thank to Tae-Kyeong Lee, a research nurse, for her valuable help.

References

AJEM

Diagnostic yield of coronary angiography in patients with acute chest pain: role of noninvasive test

a b s t r a c t

Introduction

Methods