Is Physiologic Stress Test with Imaging Comparable to Anatomic Examination of Coronary Arteries by Coronary Computed Tomography Angiography to Investigate Coronary Artery Disease? – A Systematic Review and Meta-Analysis

Objective Coronary computed tomography angiography (CCTA) is a noninvasive diagnostic modality that remains underutilized compared to functional stress testing (ST) for investigating coronary artery disease (CAD). Several patients are misdiagnosed with noncardiac chest pain (CP) that eventually die from a cardiovascular event in subsequent years. We compared CCTA to ST to investigate CP. Methods We searched MEDLINE, PubMed, Cochrane Library, and Embase from January 1, 2007 to July 1, 2018 for randomized controlled trials (RCTs) comparing CCTA to ST in patients who presented with acute or stable CP. We used Review Manager (RevMan) [Computer program] Version 5.3 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) for review and analysis. Results We included 16 RCTs enrolling 21,210 patients; there were more patients with hyperlipidemia and older patients in the ST arm compared to the CCTA arm. There was no difference in mortality: 103 in the CCTA arm vs. 110 in the ST arm (risk ratio [RR] = 0.93, 95% confidence interval [CI] = 0.71-1.21, P = .58, and I2 = 0%). A significant reduction was seen in myocardial infarctions (MIs) after CCTA compared to ST: 115 vs. 156 (RR = 0.71, CI = 0.56-0.91, P < .006, I2=0%). On subgroup analysis, the CCTA arm had fewer MIs vs. the ST with imaging subgroup (RR = 0.70, CI = 0.54-0.89, P = .004, I2 = 0%) and stable CP subgroup (RR = 0.66, CI = 0.50-0.88, P = .004, I2 = 0%). The CCTA arm showed significantly higher invasive coronary angiograms and revascularizations and significantly reduced follow-up testing and recurrent hospital visits. A trend towards increased unstable anginas was seen in the CCTA arm. Conclusions Our analysis showed a significant reduction in downstream MIs, hospital visits, and follow-up testing when CCTA is used to investigate CAD with no difference in mortality.


Introduction
Coronary heart disease is one of the leading causes of death, globally. Annually, more than 20 million patients undergo workup for angina [1]. Patients misdiagnosed with noncardiac chest pain (CP) have died from a cardiovascular event five years from the misdiagnosis [2]. Therefore, it is essential to identify patients at the highest risk of coronary artery disease (CAD) who may benefit from a workup using invasive coronary angiography (ICA) and subsequent revascularization. Coronary computed tomography angiography (CCTA) is 89% sensitive and 96% specific for the diagnosis of CAD, and CCTA is becoming an alternative to ICA due to its comparatively high diagnostic accuracy and noninvasive approach [3][4][5]. In fact, current cardiology guidelines recommend using CCTA to diagnose CAD [6].

Materials And Methods
We conducted a systematic review and meta-analysis to compare CCTA to ST with subgroup analyses of ST (with and without imaging which has never been done before) and CP (acute chest pain [ACP] or stable chest pain [SCP]). Over the years, few meta-analyses comparing CCTA to ST have been published, and the outcomes are variable; these are summarized in Table 1 [7][8][9][10][11].

FIGURE 1: PRISMA 2009 study flow diagram
PRISMA, preferred reporting items for systematic reviews and meta-analyses; RCT, randomized control trial; CCTA, coronary computed tomography angiography; ST, stress testing.

Inclusion criteria
We used the following inclusion criteria: prospective RCTs, RCTs comparing CCTA to ST after CP, age ≥ 18 years, study population ≥ 50 patients, and follow-up ≥ four weeks.
Data extraction and quality assessment W.J.S., M.S.R., and W.A. extracted data into predefined fields on a Microsoft Excel sheet for baseline characteristics and study outcomes. W.J.S. cross-checked the data and made the necessary corrections. All three reviewers discussed the revisions and agreed to the final entry.

Heterogeneity
We used I 2 statistics to calculate the heterogeneity. I 2 > 50% was considered substantial heterogeneity, as explained in the Cochrane Handbook for Systematic Reviews [29]. We performed a sensitivity analysis for considerable heterogeneity.

Results
We included 16 RCTs with 21,210 patients (10,937 in the CCTA arm and 10,273 in the ST arm). Patients in the ST arm were older than those in the CCTA arm (57.9 ± 9.8 years vs. 57.4 ± 10 years, respectively; P = .0002) and had more hyperlipidemia (45.62% vs. 43.18%, respectively; P = .0004). There was no difference in baseline body mass index, hypertension, diabetes, smoking status, and baseline use of aspirin. Three studies used ST without imaging for a total of 1,110 patients (595 in the CCTA arm and 515 in the ST without imaging arm) [17,22,26].
Complications associated with CCTA vs. ST: Only four studies reported serious complications attributed to investigation modalities used in the trials. We did not identify any difference between the two arms (7 vs. 7; RR = 0.98, CI = 0.35-2.74; P = .96, I²=0%).
Cost analysis: Eight studies reported cost, but only five studies were usable as these reported mean cost and standard deviation.
Five studies reported the total cost. There was no difference between the two arms (SMD = -0.64, CI = -1.75 to 0.46; P = .25, I²=99%). Sensitivity analysis without the CT-COMPARE reduced the heterogeneity to 45%; however, the results remained statistically insignificant. The subgroup analysis for the cost in the United States and cost elsewhere also had significant heterogeneity with no difference between the subgroups (chi-squared = 0.15, degrees of freedom = 1, P = .69, I²=0%).
Radiation dose: Four studies reported the cumulative radiation exposure usable for our analysis. The CCTA arm had significantly higher radiation exposure (SMD = 0.47, CI = 0.08-0.86; P = .02, I²=97%). Sensitivity analysis failed to reduce the heterogeneity.

Discussion
Our meta-analysis of 21,210 patients comparing CCTA to ST demonstrated a significant reduction in the primary endpoint of MIs in the CCTA group without any difference in mortality. The reduction in MI was secondary to a significantly reduced number of events in the SCP group. The reduction in MIs is likely due to the early diagnosis of obstructive CAD and subsequent early initiation of aggressive medical management and revascularizations. Recently published five-year outcomes of the SCOT-HEART trial, which enrolled patients with SCP, also showed a significant reduction in MIs over five years [30]. This discrepancy in downstream MIs between the ACP and SCP group calls for a novel assessment strategy to risk-stratify ACP patients who present to the ER regarding invasive versus conservative management. The lack of mortality benefit in our analysis may not be evident because of the short follow-up times of the individual studies (four weeks to 25 months) compared to the five-year outcomes of the SCOT-HEART trial which showed a significant reduction in mortality from coronary heart disease or nonfatal MI than standard care alone [30].
This analysis also showed increased ICA and revascularizations, which also lead to significantly reduced MIs (NNT of nine to prevent one MI for each ICA). This early difference in ICA and revascularization may be lost after an extended follow-up as suggested by the five-year outcomes of the SCOT-HEART study [23]. This indicates that CCTA use leads to early diagnosis of CAD and subsequent early intervention compared to the ST, where patients eventually needed ICA and revascularization at the cost of increased MIs and mortality. Due to the high sensitivity of CCTA (approximately 99%), a negative CCTA may reduce further testing whereas a positive CCTA leads to additional invasive procedures. In our analysis, there were significantly more ICA, true positive ICA, and revascularizations, with significantly reduced follow-up tests. The use of CCTA leads to a higher number of invasive procedures, including revascularization, ultimately leading to higher costs overall.
After the initial randomization and workup with either CCTA or ST, ER visits and rehospitalizations were significantly reduced in the CCTA arm; this differs from a previously published meta-analysis that showed no difference in ER visits and rehospitalizations [7][8][9][10][11]. A limitation of our analysis was the presence of substantial heterogeneity, making it difficult to generalize the results. The sensitivity analysis reduced heterogeneity with a trend towards reduced ER visits or rehospitalizations in the CCTA arm. Reduction in ER visits and rehospitalizations is promising, as earlier studies found that the reduced MIs after CCTA group was offset by increased future rehospitalizations and downstream costs.
The increased rates of angiographically confirmed CAD post-CCTA is another significant finding that suggests that CCTA has a better positive predictive value than ST (with or without imaging) to identify obstructive CAD at a time when current guidelines do not support the routine use of CCTA in intermediate-risk patients. Although our analysis showed an increasing trend towards unstable anginas in the CCTA arm, we hypothesize that this trend is likely the consequence of higher rates of revascularization in the CCTA group.
The cost analysis had substantial heterogeneity for both ER visits and downstream costs. The trials included in our analysis were conducted in different countries with different healthcare systems and cost structures [13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28]. In our analysis, even though a trend towards decreased ER costs was seen in the CCTA arm, there was no clear advantage of total downstream cost to either imaging strategy. In the absence of any significant mortality benefit, it is reassuring that whichever approach the provider offers will not adversely affect the patient. CCTA was associated with significantly higher cumulative radiation exposure; however, there was substantial heterogeneity, likely due to different scanners used in various trials.

Limitations
Our study had several significant limitations. First, a lack of long-term follow-up in the individual RCTs (≤25 months) that may not include events, hospitalizations, and revascularizations beyond 25 months would magnify the risks of ICA and revascularization and obscure potential long-term benefits. This may be true for ACP trials as short follow-up may have masked the advantage for either arm. Also, some outcomes were not reported by most studies, leading to substantial heterogeneity that persisted even after sensitivity analysis. In addition, we were unable to estimate radiation exposure from all studies between the two groups since they reported data in a variable form. Also, only three studies used ST without imaging, and the other studies used a combination of imaging and non-imaging ST; this leads to substantial overlap between the groups and has a risk to introduce bias in our results. Finally, these trials, although relatively modern, did not utilize high-sensitivity cardiac troponin tests. Their hypotheses must be tested again with the advent of these tests.

Conclusions
Our analysis is the largest to date of 16 RCTs and found a significant reduction in post-CCTA MIs with increased ICA and revascularizations. In the future, more RCTs are needed utilizing scoring methods to identify more robust downstream investigations, cost analysis, and radiation exposure.

Additional Information Disclosures
Human subjects: All authors have confirmed that this study did not involve human participants or tissue. Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.