Coronary computed tomography angiography (CCTA) is widely used to exclude coronary artery disease (CAD) in patients with low-to-intermediate pretest probability (PTP) of obstructive CAD. The aim of our study was to investigate the reclassification by CCTA and the implications of CCTA results on management because limited studies exist on these subjects; 1,560 patients with chest pain without a history of CAD and with low or intermediate PTP of CAD referred for CCTA from the out-patient clinic were prospectively included. PTP was defined by the Duke Clinical Score as either low (<15%), low-intermediate (15% to 50%), or high-intermediate (50% to 85%). Distribution of CCTA results among the categories of PTP of CAD and the influence of CCTA results on management were analyzed. CCTA revealed obstructive CAD in 7%, 15%, and 23% of cases, in patients with low, low-intermediate, and high-intermediate PTP, respectively; 855 of 1,031 patients (83%) with intermediate PTP of CAD showed no obstructive CAD on CCTA and were consequently reclassified. Management changes after CCTA occurred in 689 patients (44%). In 633 patients (41%), medication was altered and 135 (9%) were referred for invasive coronary angiography. Treatment with statin was initiated in 442 (28%) and stopped in 71 patients (5%). Aspirin was initiated in 192 (12%) and stopped in 139 patients (9%). In conclusion, in a routine clinical cohort, CCTA resulted in reclassification in most patients. Furthermore, our study suggests that the Duke Clinical Score overestimates the probability of obstructive CAD compared with CCTA findings. Finally, CCTA results have implications on patient management, with medication changes in 41% of patients.
Ischemic heart disease because of coronary artery disease (CAD) remains the most important cause of death worldwide. The guidelines on stable CAD recommend a stepwise approach in symptomatic patients with considered stable CAD. The process starts with the clinical assessment of the probability of obstructive CAD, the pretest probability (PTP). Risk scores such as the Duke Clinical Score are used to estimate the PTP. Noninvasive testing, such as coronary computed tomography angiography (CCTA), is advised in patients with suspected CAD. CCTA has great diagnostic performance, with a high negative predictive value. However, it is less accurate in patients with a high PTP of CAD because of overestimation of disease in these patients. Thus, the appropriate use criteria advise to use CCTA only in patients with low or intermediate PTP. In these patients, CCTA is widely used to reclassify patients by excluding or diagnosing obstructive CAD. The implications of CCTA results on patient management, however, are not clear. Therefore, the aim of this study was to describe to what extent CCTA leads to reclassification and which impact CCTA results have on patient management in out-clinic patients with chronic chest pain and low-to-intermediate PTP of obstructive CAD.
Methods
From December 13, 2011, to August 26, 2014, all patient with chronic chest pain with low or intermediate PTP of CAD referred for CCTA from the outpatient clinic were prospectively included after a diagnostic work-up. None of the patients had a history of CAD. Baseline characteristics included were prospectively entered in a database: age, gender, and cardiovascular risk factors. The PTP was calculated using the Duke Clinical Score. According to the European Society of Cardiology guidelines for stable CAD, patients with an intermediate PTP (15% to 85%) were divided into 2 groups: low-intermediate (15% to 50%) and high-intermediate (50% to 85%). Baseline and post-CCTA medication use were documented in the database. Referral for perfusion scan, invasive coronary angiography (ICA), and revascularization were ascertained from procedural reports. All patients gave written informed consent for use of their data.
Patient preparation, image acquisition, and image analysis were performed as stated subsequently and as previously described by Krul et al.
Metformin was stopped in patients with an estimated glomerular filtration rate of <60 ml/min. If heart rate at rest was >60 beats/min, 100 mg atenolol was administered orally 1 hour before CCTA. When the heart rate remained >60 beats/min, additional metoprolol intravenously could be administered in steps of 5 mg to a maximum of 30 mg. Patients were given 2 doses of nitroglycerin sublingually of 0.4 mg before CCTA.
All scans were performed with a 2 × 64-slice flying focal spot, effectively 2 × 128 slice (Somatom Definition Flash; Siemens Medical Systems, Erlangen, Germany). A 10- to 15-ml test bolus containing nonionic low-osmolar iodinated radiocontrast (ultravist 370) was injected, followed by a flush of 40 ml saline solution, both at a flow rate of 5 to 6 ml/s. The time point of maximal contrast enhancement in the ascending aorta at the level of the pulmonary trunk was recorded, and an additional delay of 5 seconds was added to define the optimal time point for acquisition of coronary artery data. A dual-head injector then injected 48 to 75 ml contrast depending on kilovolts used for the high-pitch flash scan and 75 ml in case of prospective or retrospectively triggered scan, followed by 45-ml 30/70% contrast/saline solution at a flow rate of 6 ml/s. The tube voltage (80, 100, or 120 kV) and tube current were determined automatically by the scanning system based on body geometry. The delivered radiation dose was also generated automatically by the scanner software and represented as dose length product (DLP). The effective dose of the scan was calculated by multiplying the DLP with the k-factor of 0.014 mSv (mGy × cm) −1 , which is generally used for effective dose estimation in cardiac CT studies. The total dose of topogram, test bolus, coronary artery calcium score, and CCTA was used to estimate the effective radiation dose for each patient. All scans were read in consensus by a Certification Board Of Cardiovascular Computed Tomography accredited nuclear medicine physician and a cardiologist experienced in the interpretation of CCTA. In case of disagreement, a third opinion was decisive. For the coronary artery calcium Score, coronary calcifications were defined as dense lesions in coronary arteries with densities >130 Hounsfield units. Calcifications were manually assigned to coronary arteries and added to the Agatston score for each patient.
Structures >1 mm 2 within and/or adjacent to the coronary artery lumen, which could be clearly distinguished from the vessel lumen, were scored as a coronary plaque. Coronary plaques were scored per coronary segment. Each coronary plaque was quantified for stenosis by visual estimation.
Obstructive CAD was defined as a lumen stenosis in any of the major coronary vessels of >50%, either left main artery, left anterior descending artery, circumflex artery, or right coronary artery. Normal coronary arteries were defined as CAC score = 0 and no coronary plaque. Nonobstructive CAD was defined as CAC score >0 and/or any plaque that did not meet the criteria for obstructive CAD. In obstructive CAD, the extent of CAD was recorded, 1-, 2-, or 3-vessel CAD. High-risk anatomic features were defined as left main artery stenosis >50%, >50% stenosis in the proximal left anterior descending artery, or 3-vessel CAD with >50% stenosis.
Statistical analysis was performed using SPSS software, version 22.0.0 (SPSS Inc, Chicago, Illinois). Continuous variables are presented as mean ± SD and categorical variables as frequencies with percentages. Variables were compared with the chi-square test for categorical variables and by applying an analysis of variance or an unpaired student t test for continuous variables as appropriate. Continuous variables were tested for normal distribution.
The presence of CAD at CCTA was evaluated in the different categories of the Duke Clinical Score—low, low-intermediate, and high-intermediate—using descriptive statistics. Descriptive statistics were also used to evaluate the management changes in patients with normal coronary arteries, patients with nonobstructive CAD, and patients with obstructive CAD. Changes in patient management consisted of 2 different types of changes: changes in cardiac medication and/or referral for ICA. Cardiac medication changes between the baseline situation and after CCTA were defined as discontinuation or addition of aspirin or statin. The referral for perfusion scans, either 13 NH 3 positron emission tomography or single-photon emission computed tomography, was evaluated in patients with obstructive CAD and those with obstructive CAD and high-risk anatomic features. Referral for ICA was defined as the occurrence of ICA in the first 60 days after CCTA. As in previous studies, this 60-day landmark was used to differentiate between CCTA-driven ICA and long-term revascularization, which is considered to be indicative for the prognosis. Acute revascularizations in myocardial infarction patients were also considered non-CCTA driven. The 60-day landmark was also used for perfusion scans.
Results
The baseline characteristics of the total study population (n = 1,560) and the baseline characteristics in the different categories of the PTP are summarized in Table 1 . Cardiac risk factors and baseline medication use were more frequent in higher PTP categories.
| Variable | Total (n=1560) | Pre-test probability | p-value ∗ | ||
|---|---|---|---|---|---|
| Low (n=529) | Low- intermediate (n=764) | High- intermediate (n=267) | |||
| Demographics | |||||
| Age (years) | 57.9 ± 10.2 | 54.9 ± 10.1 | 58.6 ± 10.2 | 62.1 ± 8.8 | <0.001 |
| Women | 971 (62.2%) | 462 (87.3%) | 424 (55.5%) | 85 (31.8%) | <0.001 |
| Body Mass Index (kg/m 2 ) | 26.6 ± 4.5 | 26.5 ± 5.1 | 26.7 ± 4.3 | 26.8 ± 3.8 | 0.58 |
| Diabetes Mellitus | 122 (7.8%) | 24 (4.5%) | 64 (8.4%) | 34 (12.8%) | 0.001 |
| Hba1c (n=53) † | 6.8 ± 1.6 | ||||
| Hypertension | 464 (29.7%) | 134 (25.3%) | 244 (31.9%) | 86 (32.2%) | 0.024 |
| Hyperlipidemia | 396 (25.4%) | 114 (21.6%) | 208 (27.2%) | 74 (27.7%) | 0.044 |
| Family history of CAD ‡ | 738 (47.4%) | 258 (48.9%) | 362 (47.4%) | 118 (44.2%) | 0.46 |
| Smoker | 279 (17.9%) | 77 (14.6%) | 152 (19.9%) | 50 (18.7%) | 0.045 |
| eGFR<60 | 44 (2.8%) | 15 (2.8%) | 24 (3.1%) | 5 (1.9%) | 0.56 |
| Baseline medication | |||||
| Aspirin | 501 (32.1%) | 131 (24.8%) | 243 (31.8%) | 127 (47.6%) | <0.001 |
| Statin | 534 (34.2%) | 122 (23.1%) | 277 (36.3%) | 135 (50.6%) | <0.001 |
| Beta-blocker | 606 (38.8%) | 166 (35.2%) | 286 (37.4%) | 134 (50.2%) | <0.001 |
| ACE-i / ARB | 392 (25.1%) | 108 (20.4%) | 203 (26.6%) | 81 (30.3%) | 0.004 |
| Calcium channel blocker | 96 (6.2%) | 32 (6.0%) | 50 (6.5%) | 14 (5.2%) | 0.74 |
| Nitrate | 49 (3.1%) | 6 (1.1%) | 24 (3.1%) | 19 (7.1%) | <0.001 |
| Acenocoumarol | 38 (2.4%) | 9 (1.7%) | 17 (2.2%) | 12 (4.5%) | 0.047 |
∗ p-value for distribution of baseline characteristics among the different categories of the pre-test probability.
† Hba1C was documented in only 53 patients.
The CCTA results are listed in Table 2 . In 48 patients, CCTA showed obstructive CAD with high-risk anatomic features. Two patients with left main CAD on CCTA also had 3-vessel disease. The effective radiation delivered to the entire group was 2.4 ± 2.0 mSv (DLP 174 ± 142; mean ± SD). The high-pitch flash scans (n = 1,130) resulted in an effective dose of 1.6 ± 0.7 mSv (DLP 114 ± 47), the prospectively triggered scans (n = 395) in 4.5 ± 2.45 mSv (DLP 322 ± 176), and the retrospectively triggered scans (n = 35) in 6.5 ± 2.8 mSv (DLP 463 ± 196).
| Variable | Total (n=1560) | Pre-test probability | ||
|---|---|---|---|---|
| Low PTP (n=529) | Low- intermediate PTP (n=764) | High- intermediate PTP (n=267) | ||
| Calcium (Agatston score) | ||||
| 0 | 744 (47.7%) | 336 (63.5%) | 334 (43.7%) | 74 (27.7%) |
| 0.1-100 | 501 (32.1%) | 127 (24.0%) | 268 (35.1%) | 106 (39.7%) |
| 100-400 | 216 (13.8%) | 54 (10.2%) | 113 (14.8%) | 49 (18.4%) |
| >400 | 99 (6.3%) | 12 (2.3%) | 49 (6.4%) | 38 (14.2%) |
| CCTA results | ||||
| Normal coronary arteries | 659 (42.2%) | 308 (58.2%) | 294 (38.5%) | 57 (21.3%) |
| Non-obstructive CAD | 686 (44.0%) | 182 (34.4%) | 354 (46.3%) | 150 (56.2%) |
| Number of coronary arteries narrowed (>50%) | 215 (13.8%) | 39 (7.4%) | 116 (15.2%) | 60 (22.5%) |
| 1 | 165 (10.6%) | 31 (5.9%) | 94 (12.3%) | 39 (14.6%) |
| 2 | 34 (2.2%) | 6 (1.1%) | 13 (1.7%) | 15 (5,6%) |
| 3 | 13 (0.8%) | 1 (0.2%) | 7 (0.9%) | 5 (1.9%) |
| Left main | 4 (0.3%) | 1 (0.2%) | 2 (0.3%) | 1 (0.4%) |
| High risk narrowings ∗ | 48 (3.1%) | 5 (0.9%) | 30 (3.9%) | 13 (4.9%) |
∗ High risk narrowings were defined as left main, 3-vessel and/or proximal LAD disease.
CCTA led to reclassification in 894 patients (57%), of which 39 of 529 (7%) of the low PTP category were reclassified as having obstructive CAD, whereas 855 of 1,031 (83%) of the low-intermediate and high-intermediate PTP category without obstructive CAD were reclassified as having no obstructive CAD. The reclassification is illustrated in Figure 1 .
The alteration in patient management after CCTA is listed in Table 3 and Figure 2 . Presence of obstructive CAD, and nonobstructive CAD, was associated with medication changes and referral for ICA (p <0.001). Patient management was changed in 689 patients (44%) after CCTA. Alteration of baseline medication after CCTA occurred in 633 patients (41%). Treatment with statin was initiated in 442 patients (28%) and stopped in 71 patients (5%). In 192 patients (12%), treatment with aspirin was initiated and in 139 (9%) aspirin was stopped after CCTA. Of all patients with normal coronary arteries, aspirin was discontinued in 105 patients after CCTA (16%).
