You are here

Article Text

Article menu
Coronary artery disease
Original research
Outcomes in patients with a first episode of chest pain undergoing early coronary CT imaging
  1. Klaske R Siegersma1,2,
  2. N Charlotte Onland-Moret3,
  3. Yolande Appelman1,
  4. Pim van der Harst4,
  5. Igor I Tulevski5,
  6. G Aernout Somsen5,
  7. Jagat Narula6,
  8. Hester M den Ruijter2,
  9. Leonard Hofstra1,5
  1. 1 Department of Cardiology, Amsterdam UMC VUMC Site, Amsterdam, The Netherlands
  2. 2 Laboratory of Experimental Cardiology, University Medical Center Utrecht, Utrecht University, Utrecht, The Netherlands
  3. 3 Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht University, Utrecht, The Netherlands
  4. 4 Department of Cardiology, University Medical Center Utrecht, Utrecht University, Utrecht, The Netherlands
  5. 5 Cardiology Centers of the Netherlands, Utrecht, The Netherlands
  6. 6 Icahn School of Medicine at Mount Sinai, New York, New York, USA
  1. Correspondence to Professor Leonard Hofstra, Cardiology Centers of the Netherlands, Utrecht, 3584 AA, The Netherlands; leonard.hofstra{at}gmail.com

Abstract

Objectives To investigate the impact of a CT-first strategy on all-cause and cardiovascular mortality in patients presenting with chest pain in outpatient cardiology clinics.

Methods Patients with a first presentation of suspected angina pectoris were identified and their data linked to the registrations of Statistics Netherlands for information on mortality. The linked database consisted of 33 068 patients. CT-first patients were defined as patients with a CT calcium score and coronary CT angiography, within 6 weeks after their initial visit. Propensity score matching (1:5) was used to match patients with and without a CT-first strategy. After matching, 12 545 patients were included of which 2308 CT-first patients and 10 237 patients that underwent usual care.

Results Mean age was 57 years, 56.3% were women and median follow-up was 4.9 years. All-cause mortality was significantly lower in CT-first patients (n=43, 1.9%) compared with patients without CT (n=363, 3.5%) (HR: 0.51, 95% CI 0.37 to 0.70). Furthermore, CT-first patients were more likely to receive cardiovascular preventative and antianginal medication (aspirin: 44.9% vs 27.1%, statins: 48.7% vs 30.3%, beta-blockers: 37.8% vs 25.5%, in CT-first and without CT-first patients, respectively) and to undergo downstream diagnostics and interventions (coronary interventions: 8.5% vs 5.7%, coronary angiography: 16.2% vs 10.6% in CT-first and without CT-first patients, respectively).

Conclusions In a real-world regular care database, a CT-first strategy in patients suspected of angina pectoris was associated with a lowering of all-cause mortality.

  • coronary artery disease
  • computed tomography angiography
  • diagnostic imaging
  • electronic health records

Data availability statement

Data are available on reasonable request. The CCN data is not publicly available due to ethical and data protection constraints, but are available from the corresponding author on reasonable request. The results of presented survival analyses are based on calculations by the authors using non-public microdata from Statistics Netherlands. Under certain conditions, these microdata are accessible for statistical and scientific research. For further information: microdata@cbs.nl.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/heartjnl-2021-319747

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Longitudinal studies of employing cardiac CT imaging including both coronary calcium scoring (CACS) and cardiac CT angiography (CCTA) in patients presenting with chest pain have demonstrated incremental prognostic value compared with traditional risk profiling algorithms.1 Higher CACS is associated with mortality2 3 and has added value to the Framingham Risk Score for predicting cardiovascular events.4 The introduction of CCTA showed even greater promise as a tool to define risk of myocardial infarction5 and coronary revascularisation6 compared with CACS. Prospectively randomised trials showed that CCTA was superior to functional cardiac testing for cardiovascular endpoints.7 8

    Until the results of the SCOT-HEART (Scottish COmputed Tomography of the HEART Trial) study were published,5 CCTA was only adopted by the guidelines of the National Institute for Health and Care Excellence in the UK.9 The SCOT-HEART study randomised patients presenting with chest pain in the cardiology outpatient clinic to either CCTA initiated, the so-called CT-first strategy, or routine clinical care. Patients randomised to the CCTA arm showed only about half of the fatal and non-fatal myocardial infarctions after a 5-year follow-up,5 although these results were not in line with the previously published PROMISE-trial8 and multiple registry studies.10 11 Nevertheless, these results prompted the European Society of Cardiology to accord CCTA a class I recommendation for the diagnosis of coronary artery disease (CAD) in symptomatic patients in the renewed European Society of Cardiology (ESC) guidelines of 2019.12

    However, the results of SCOT-HEART might not be generalisable to a routine care population who are expected to have a higher likelihood of mortality, since randomised trials may induce an overestimation of benefit due to the so-called healthy volunteer inclusion bias in trials.13 14 Furthermore, multiple patient groups are often under-represented in clinical studies, such as women,15 the elderly, non-white ethnicities and patients with comorbidities.16 17 Therefore, the impact of a CT-first strategy for patients with chest pain in regular care is still unknown.10 11

    The use of data from real-world databases to assess the utility of cardiac CT may aid in defining associations between patient variables and outcomes in the general population. Moreover, real-world data may help to vindicate results of clinical trials, especially with the inclusion of substantial numbers of all relevant patient groups. Thus, we used a regular care database18 to investigate all-cause and cardiovascular mortality in patients suspected of angina pectoris, who underwent CT calcium score and CCTA following their first visit, and compared this with a propensity score (PS) matched control group from the same database, who were subjected to regular care without cardiac CT.

    Methods

    Study population

    The CCN database consists of 109 151 patients who visited one of the CCN’s outpatient cardiology clinics between 2007 and 2018. These centres are known for their homogenised and structured approach to investigate patients with cardiovascular complaints. A detailed description of the CCN database has been published.18 Patients with chest pain suspected to be angina pectoris, who presented for the first time at one of the diagnostic centres, were included. The study population was then split in two groups: patients with a cardiac CT within 6 weeks after their first visit, the CT-first strategy, and patients without this diagnostic procedure. Cardiac CT included a scanning protocol for calcium scoring and CCTA. The decision for referral for cardiac CT was made by the treating cardiologist. The type of CT scanner was based on availability in the referenced centres. Follow-up of the population was obtained through linking with the population database of Statistics Netherlands (CBS). Figure 1 outlines the selection of the study population.

    Figure 1
    Figure 1

    Flow chart of patient selection. CCN, Cardiology Centers of the Netherlands.

    The Cardiology Centers of the Netherlands data were made available under implied consent and transferred to the University Medical Center Utrecht under the Dutch Personal Data Protection Act. This study used data collected during the regular care process and did not subject participants to additional procedures or impose behavioural patterns on them. The Medical Research Ethics Committee of the University Medical Center Utrecht declared that research within the CCN database does not meet the Dutch Medical Research Involving Human Subjects Act (proposal number 17/359).

    Study variables

    Patient characteristics, prior comorbidities, risk factors, family history and a general medical history were obtained from the electronic health records (EHRs) of the patient. Medication use was extracted from pharmaceutical prescription data. Residential region of the patient was obtained through four-digit postal code. Chest pain was characterised as typical, atypical or non-anginal, according to Diamond12 19 and obtained through retrieval of text variables in the EHR. We also included the results from the stress ECG recording to ensure comparability between patients with and without a CT-first strategy.

    Short-term and long-term outcome

    All-cause mortality and cardiovascular mortality were the primary and secondary outcome in the analyses, respectively. Other secondary outcomes included registration of diagnostic and therapeutic procedures in the EHR. Diagnostics included stress ECG recording, functional imaging (cardiac MRI, positron emission tomography and single-photon emission computed tomography), invasive coronary angiography (CAG) and cardiac CT more than 6 weeks after the first visit to CCN. Therapeutic procedures were percutaneous coronary intervention and coronary artery bypass grafting. Moreover, prescribed medication changes during and after the chest pain consult were evaluated and compared between the groups. This analysis focused primarily on the value of a cardiac CT in the diagnostic trajectory of the patient. We did not evaluate the association between results of cardiac CT and primary and secondary outcomes.

    Missing data

    Missing data in variables required for analysis were handled by multiple imputation for chained equations (MICE; 10 iterations, 10 imputed datasets) with the R-package mice (V.3.8.0).20 However, for the presence of a family history, missing data were not at random. Therefore, missing values for this variable were filled with a negative family history. Included variables for multiple imputation models are indicated in table 1.

    View this table:
    Table 1

    Baseline table of total population and PS matched sample

    PS matching

    PS matching21 was used to ensure a comparable sample between patients with and without a CT-first strategy. PSs were calculated for each patient separately in each imputed dataset with a logistic regression model including 19 variables, as indicated in table 1. This selection of variables was based on clinically relevant variables and baseline differences between both groups. Thereafter, the average PS of the imputation datasets was calculated for each patient. We matched five patients who did not receive a CT-first strategy to one patient with a CT-first strategy based on the calculated PSs per patient using the nearest neighbour method with a calliper width of 0.05 and no replacement (R-package: matchit,22 V.3.0.2). Comparability of the groups after matching was assessed by inspection of the balance of baseline variables. The selected patient population is further referred to as the matched sample.

    Statistical analysis

    Descriptive statistics are presented as mean with SD or median with IQR, where appropriate, for continuous variables and counts and percentages for categorical variables. All-cause and cardiovascular mortality were analysed with Kaplan-Meier curves and Cox regression models on the matched patient selection. Other secondary outcomes were compared with χ2 testing.

    Subgroup analyses were performed for type of chest pain (anginal/non-anginal), sex (men/women), age (<65 years/≥65 years), Systematic COronary Risk Evaluation (SCORE) (<5%/≥5%)23 and pretest probability of CAD (<5%/>5%).24 The p value for interaction was determined for each analysis. Outcome for all subgroup analyses was all-cause mortality.

    Sensitivity analyses were done to evaluate the effect of patient’s residency and inclusion year. In these analyses, a Cox regression for all-cause mortality was performed. Another sensitivity analysis excluded all patients referred to one specific diagnostic centre, only performing CCTA in case of medical need or if calcium score was above 0. As no information of the centre that was visited is available in the database, residency of the patient is used as a substitute. A final sensitivity analysis removed all patients without a CT but with CAG to evaluate the effect on the results. These sensitivity analyses evaluated Cox regressions for both cardiovascular and all-cause mortality.

    All data analyses were done with R (V.3.6.2) and RStudio (V.1.1.463).

    Patient and public involvement

    Patients were not involved in any stage of this research process.

    Results

    Study population

    A total of 34 311 patients with chest pain met the inclusion criteria and were selected from the CCN database (figure 1). After linking with the database of Statistics Netherlands, mortality data of 33 068 (96.4%) patients were available. This selection included 2312 and 30 756 patients, respectively, with and without a CT-first strategy.

    Mean age of the included patients (n=33 068) was 56 years and 53.3% were women. Before PS matching, patients in the CT-first group (table 1) were older (57 vs 55 years), less likely to be current smokers (35.5% vs 42.6%) and had a higher incidence of typical chest pain (45.7% vs 27.1 %). Yet, the presence of comorbidities was comparable. Median follow-up was 5.5 (IQR: 3.4–7.8) years, and 1331 (4.0%) patients died during follow-up. In 329 (1.0%) patients, a cardiovascular cause of death was assigned. Thus, PSs were distributed differently between patients with and without a CT-first strategy (online supplemental figure 1).

    Supplemental material

    After PS matching, 2308 patients were in the CT-first group, and 10 237 patients were matched to this group. Patients in the CT-first group had a median coronary artery calcification score of 2.20 (IQR: 0–88). There were 871, 329 and 225 filed stenoses in the left anterior descending artery, right coronary artery and circumflex artery, respectively, of which 489 (34.3%) were reported to be significant (>50% stenosis). In the matched sample, baseline characteristics were more equally distributed as compared with the unmatched sample (table 1), for example, mean age (57 years in both groups) and type of chest pain (non-anginal 39.1% vs 36.1%, atypical 15.3% vs 16.3%, and typical 45.6% vs 47.6% for, respectively, with and without a CT-first strategy). Consequently, the distribution of the PS of the included patients showed improved overlap in the matched sample (online supplemental figure 1). The matched sample (n=12 545) was used for the analyses of primary and secondary outcomes.

    Outcome analysis

    All-cause and cardiovascular mortality after PS matching

    In the matched sample 406 (3.2%) patients died. In 111 (0.9%) patients, there was a cardiovascular cause of death. Median follow-up was 4.9 years. In patients with a CT-first strategy, all-cause mortality was 1.9% (n=43) compared with 3.5% (n=363) in patients without a CT-first strategy (HR: 0.51, 95% CI 0.37 to 0.70). The Kaplan-Meier curves for patients with and without a CT-first strategy are shown in figure 2 and show a consistent divergent pattern over up to 7 years of follow-up. Cardiovascular mortality for patients in the CT-first group was 16 (0.7%) and 95 (0.9%) for the patients without a CT-first strategy, respectively. The corresponding HR for cardiovascular mortality was 0.73 (95% CI 0.43 to 1.24). The corresponding Kaplan-Meier curves are shown in online supplemental figure 2.

    Figure 2
    Figure 2

    Kaplan-Meier curves for all-cause mortality of the CT-first and the without a CT-first study population.

    Downstream diagnostics and interventions

    A total of 3432 (27.4%) patients of the matched cohort had diagnostic or therapeutic follow-up. The percentage of CAGs (16.2% vs 10.6%, p<0.001) and coronary interventions (8.5% vs 5.7%, p<0.001) were higher in the CT-first strategy group, compared with the group of patients without CT (table 2). The same was seen for the number of stress ECGs, which was higher for CT-first patients (17.5% vs 15.6%, p=0.03), yet the time interval between initial presentation and stress ECG was shorter in patients without a CT-first strategy (median time between chest pain consult and stress ECG: 112 days, IQR: 25–598, vs 301 days, IQR: 91–759). Follow-up cardiac perfusion imaging did not show any differences (3.0% vs 2.8% in, respectively, patients with and without a CT-first strategy).

    View this table:
    Table 2

    Distribution of diagnostics and therapeutics during follow-up

    Medication use

    Medication use for primary or secondary prevention of CAD (including aspirin, statins, beta-blockers, calcium channel blockers and nitrates) at baseline was similar for patients with and without a CT-first strategy (tables 1 and 3). After initial chest pain consult, initiation of aspirin (44.9% vs 27.1%, p<0.001), b-blockers (37.8% vs 25.5%, p<0.001) and statins (48.7% vs 30.3%, p<0.001) was higher in the CT-first strategy group, compared with the group without CT.

    View this table:
    Table 3

    Medication use in selected population

    Subgroup and sensitivity analyses

    The association between having a cardiac CT for diagnostic evaluation of chest pain and all-cause mortality was further investigated in clinically relevant subgroups (figure 3). Men showed greater benefit from a CT-first strategy with respect to all-cause mortality, although p value for interaction was not significant. This p-value was only significant for type of chest pain (p-value: 0.013) and risk of CAD (p-value: 0.0046), implying that patients with a possible anginal origin of chest pain and with an intermediate to high risk of CAD benefit from a CT-first strategy.

    Figure 3
    Figure 3

    Forest plot of the subgroup analysis. CAD risk is defined as the pretest probability of CAD according to the ESC guidelines of 2019: Low : <5%, intermediate to high: >5%. Anginal chest pain includes typical and atypical chest pain. SCORE is the cardiovascular risk score as defined by the Systematic COronary Risk Evaluation. CAD, coronary artery disease; ESC, European Society of Cardiology.

    Sensitivity analyses showed that HRs were similar for the different regions of the Netherlands and for the inclusion year of patients (online supplemental table 1). Residence and inclusion year were taken into account during the PS calculation. This suggests a similar distribution of these variables in the CT-first and without CT-first group. Sensitivity analysis of removal of patients (n=4883) from one centre that did not structurally perform cardiac CT including calcium scoring and CCTA showed comparable HR for all-cause mortality (0.44, 95% CI 0.28 to 0.67 vs 0.51, 95% CI 0.37 to 0.70) and HR for cardiovascular mortality (0.70, 95% CI 0.37 to 1.33 vs 0.73 95% CI 0.43 to 1.23) to the main matched sample. After removal of all patients with CAG within 10 weeks after their initial visit (n=959), the HR for all-cause mortality (0.54, 95% CI 0.39 to 0.74) and cardiovascular mortality (0.78, 95% CI 0.45 to 1.34) changed but within CIs.

    Discussion

    The analysis performed in this study demonstrated that upstream inclusion of cardiac CT in diagnostic work-up of patients with chest pain was associated with a significant lowering of all-cause mortality, irrespective of the outcome of the cardiac CT. We also observed a lower cardiovascular mortality rate in CT-first patients, although this difference was not statistically significant. This can be due to the low numbers of a cardiovascular cause of deaths and insufficient registration of the cause of death. Our data also show that patients with a CT-first strategy had more downstream therapeutics and diagnostic testing. This is in accordance with findings from other registries.10 The presented results underscore and complement the prominent role of cardiac CT in the diagnostic work-up of patients presenting with chest pain in clinical guidelines,12 which has also been internationally addressed.25

    Based on previous publications, it is assumed that patients that undergo cardiac CT in addition to regular care receive an accurate diagnosis earlier and more often,26 after which targeted antianginal and preventative therapy is initiated, resulting in overall superior survival in these patients.27 28 The prescription of preventative medication, especially statins, aspirin and beta blockers to patients with CT-verified atherosclerotic lesions was thought to be the major reason for lower mortality rates observed in patients undergoing cardiac CT in the SCOT-HEART trial. However, cardiovascular mortality rates were not significantly lower in the CT-first population. One reason for the lower risk of all-cause mortality could be the identification of relevant extracardiac findings, which occur in approximately 3%. Of these, pulmonary nodules make up the largest share (1.3%).29 Also the inclusion of patients at lower risk for CAD in the CT-first group that have not been properly identified due to missing values might account for the discrepancy in all-cause and cardiovascular mortality after a CT-first strategy. Finally, the number of events and follow-up time might be too low to establish the effect of lifestyle changes as a result of the visualisation of coronaries on CT. Yet, as we did not include an in-depth analysis of the results of cardiac CT, we could not relate this to the presence of CAD or atherosclerotic lesions.

    In accordance with the SCOT-HEART trial, we observed a higher referral rate for follow-up diagnostics and interventions in CT-first patients.5 27 We did not perform analyses of these downstream costs. Therefore, no statements can be given about the cost-to-benefit ratio in patients with and without a cardiac CT-first strategy. Yet, in the light of a faster and more often correct diagnosis,27 lower cumulative diagnostic expenses could be expected for patients with a cardiac CT.26 Nevertheless, cost-effectiveness of the CT-first strategy needs to be established and should take into account the higher number of revascularisations and medication prescriptions in CT-first patients.10 30 On the contrary, the relatively low cost and wide availability of CT makes it accessible for more patients compared with other cardiac imaging strategies.

    Strengths and limitations

    The presented study has multiple strengths. First, the use of EHR data aids to address concerns of health volunteer bias and under-representation of patient subgroups that are mostly seen in clinical trials. This is represented in the large study population and equal distribution of sex within the population. Thus, the included population is an actual representation of the population that the guidelines are intended for. Therefore, the results have high external validity. Second, the longer follow-up in this study has incremental value compared with the Danish nationwide registry10 and the PROMISE trial8 that presented data with a follow-up of respectively 3.6 years and 25 months.

    The study was limited by the lack of information on cardiovascular events, including cardiovascular hospitalisations, which hampered the analysis of these events. Another limitation of regular care data is confounding by indication; patients with a higher pretest probability of CAD are more likely to be referred for cardiac CT than patients with a lower pretest probability. To avoid this potential bias, PS matching was used to ensure comparability between patients with and without a CT-first strategy. Regardless, one would have expected that patients with a higher pretest likelihood would show lower survival. Yet, we observed the opposite, vindicating that baseline differences between both groups were properly accounted for. Nonetheless, it is impossible to eliminate or to take into account the influence of unmeasured confounders or instrumental variables in the presented methodology.

    To conclude, this study was the first to demonstrate a significantly lower HR for all-cause mortality in patients with chest pain who had a CT-first strategy compared with patients who did not have a CT-first strategy. It is hypothesised that patients with a CT-first strategy obtain a more tailored therapy, including risk-reducing medication. These results support a CT-first strategy for patients with chest pain and strengthen the prominent role for cardiac CT as the primary method for diagnostic work-up of patients with chest pain, as suggested by the current ESC guidelines.

    Key messages

    What is already known on this subject?

    • Upstream incorporation of cardiac CT in the diagnostic pathway (CT-first strategy) has shown to reduce fatal and non-fatal myocardial infarction with 50% in patients with chest pain in the SCOT-HEART study. Although registry studies and the PROMISE trial could not confirm these findings, international guidelines have incorporated a CT-first strategy in the evaluation of patients with chest pain. Regular care data can aid in building the evidence for this guideline implementation.

    What might this study add?

    • A CT-first strategy in patients with chest pain is associated with a lower risk of all-cause mortality (HR: 0.51, 95% CI 0.37 to 0.70), but not cardiovascular mortality (HR: 0.73, 95% CI 0.43 to 1.24). In patients with a CT-first strategy, prescriptions of risk-reducing and antianginal medication were higher compared with patients without a CT-first strategy.

    • Propensity score matching and multiple imputation can reduce bias present in regular care data. These methods enable researchers to use the potential of regular care data to strengthen the evidence base for clinical practices.

    How might this impact on clinical practice?

    • Patients undergoing a cardiac CT as part of their diagnostic work-up for chest pain might have increased survival, compared with patients undergoing usual care. It is hypothesised this is due to more timely and accurate diagnosis of the origin of chest pain, resulting in targeted therapeutics and initiation of antianginal and preventative medication.

    Data availability statement

    Data are available on reasonable request. The CCN data is not publicly available due to ethical and data protection constraints, but are available from the corresponding author on reasonable request. The results of presented survival analyses are based on calculations by the authors using non-public microdata from Statistics Netherlands. Under certain conditions, these microdata are accessible for statistical and scientific research. For further information: microdata@cbs.nl.

    Ethics statements

    Patient consent for publication

    Acknowledgments

    Results are based on calculations by the authors using non-public microdata from Statistics Netherlands.

    References

    1. ↵
      2. Arad Y ,
      3. Spadaro LA ,
      4. Roth M , et al
      . Treatment of asymptomatic adults with elevated coronary calcium scores with atorvastatin, vitamin C, and vitamin E: the St. Francis heart study randomized clinical trial. J Am Coll Cardiol 2005;46:166–72.doi:10.1016/j.jacc.2005.02.089 pmid:http://www.ncbi.nlm.nih.gov/pubmed/15992652
      OpenUrlFREE Full Text
    2. ↵
      2. Budoff MJ ,
      3. Shaw LJ ,
      4. Liu ST , et al
      . Long-Term prognosis associated with coronary calcification: observations from a Registry of 25,253 patients. J Am Coll Cardiol 2007;49:1860–70.doi:10.1016/j.jacc.2006.10.079 pmid:http://www.ncbi.nlm.nih.gov/pubmed/17481445
      OpenUrlFREE Full Text
    3. ↵
      2. Detrano R ,
      3. Guerci AD ,
      4. Carr JJ , et al
      . Coronary calcium as a predictor of coronary events in four racial or ethnic groups. N Engl J Med 2008;358:1336–45.doi:10.1056/NEJMoa072100 pmid:http://www.ncbi.nlm.nih.gov/pubmed/18367736
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      2. Yeboah J ,
      3. McClelland RL ,
      4. Polonsky TS , et al
      . Comparison of novel risk markers for improvement in cardiovascular risk assessment in intermediate-risk individuals. JAMA 2012;308:788–95.doi:10.1001/jama.2012.9624 pmid:http://www.ncbi.nlm.nih.gov/pubmed/22910756
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. SCOT-HEART Investigators,
      2. Newby DE ,
      3. Adamson PD , et al
      . Coronary CT angiography and 5-year risk of myocardial infarction. N Engl J Med 2018;379:924–33.doi:10.1056/NEJMoa1805971 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30145934
      OpenUrlCrossRefPubMed
    6. ↵
      2. Villines TC ,
      3. Hulten EA ,
      4. Shaw LJ , et al
      . Prevalence and severity of coronary artery disease and adverse events among symptomatic patients with coronary artery calcification scores of zero undergoing coronary computed tomography angiography: results from the CONFIRM (coronary CT angiography evaluation for clinical outcomes: an international multicenter) registry. J Am Coll Cardiol 2011;58:2533–40.doi:10.1016/j.jacc.2011.10.851 pmid:http://www.ncbi.nlm.nih.gov/pubmed/22079127
      OpenUrlFREE Full Text
    7. ↵
      2. Neglia D ,
      3. Rovai D ,
      4. Caselli C , et al
      . Detection of significant coronary artery disease by noninvasive anatomical and functional imaging. Circ Cardiovasc Imaging 2015;8:1–10.doi:10.1161/CIRCIMAGING.114.002179 pmid:http://www.ncbi.nlm.nih.gov/pubmed/25711274
      OpenUrlCrossRefPubMed
    8. ↵
      2. Douglas PS ,
      3. Hoffmann U ,
      4. Patel MR , et al
      . Outcomes of anatomical versus functional testing for coronary artery disease. N Engl J Med 2015;372:1291–300.doi:10.1056/NEJMoa1415516 pmid:http://www.ncbi.nlm.nih.gov/pubmed/25773919
      OpenUrlCrossRefPubMed
    9. ↵
      1. National Institute for Health and Clinical Excellence
      . Recent-onset chest pain of suspected cardiac origin: assessment and diagnosis. Clin Guidel [CG95] 2010.
    10. ↵
      2. Jørgensen ME ,
      3. Andersson C ,
      4. Nørgaard BL , et al
      . Functional Testing or Coronary Computed Tomography Angiography in Patients With Stable Coronary Artery Disease. J Am Coll Cardiol 2017;69:1761–70.doi:10.1016/j.jacc.2017.01.046 pmid:http://www.ncbi.nlm.nih.gov/pubmed/28385304
      OpenUrlFREE Full Text
    11. ↵
      2. Roifman I ,
      3. Wijeysundera HC ,
      4. Austin PC , et al
      . Comparison of anatomic and clinical outcomes in patients undergoing alternative initial noninvasive testing strategies for the diagnosis of stable coronary artery disease. J Am Heart Assoc 2017;6:1–13.doi:10.1161/JAHA.116.005462
      OpenUrlCrossRefPubMed
    12. ↵
      2. Knuuti J ,
      3. Wijns W ,
      4. Saraste A , et al
      . 2019 ESC guidelines for the diagnosis and management of chronic coronary syndromes. Eur Heart J 2020;41:407–77.doi:10.1093/eurheartj/ehz425 pmid:http://www.ncbi.nlm.nih.gov/pubmed/31504439
      OpenUrlCrossRefPubMed
    13. ↵
      2. Pinsky PF ,
      3. Miller A ,
      4. Kramer BS , et al
      . Evidence of a healthy volunteer effect in the prostate, lung, colorectal, and ovarian cancer screening trial. Am J Epidemiol 2007;165:874–81.doi:10.1093/aje/kwk075 pmid:http://www.ncbi.nlm.nih.gov/pubmed/17244633
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      2. Leening MJG ,
      3. Heeringa J ,
      4. Deckers JW , et al
      . Healthy volunteer effect and cardiovascular risk. Epidemiology 2014;25:470–1.doi:10.1097/EDE.0000000000000091 pmid:http://www.ncbi.nlm.nih.gov/pubmed/24713887
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      2. Pilote L ,
      3. Raparelli V
      . Participation of women in clinical trials: not yet time to rest on our laurels. J Am Coll Cardiol 2018;71:1970–2.doi:10.1016/j.jacc.2018.02.069 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29724349
      OpenUrlFREE Full Text
    16. ↵
      2. Sardar MR ,
      3. Badri M ,
      4. Prince CT , et al
      . Underrepresentation of women, elderly patients, and racial minorities in the randomized trials used for cardiovascular guidelines. JAMA Intern Med 2014;174:1868–70.doi:10.1001/jamainternmed.2014.4758 pmid:http://www.ncbi.nlm.nih.gov/pubmed/25264856
      OpenUrlPubMed
    17. ↵
      2. Van Spall HGC ,
      3. Toren A ,
      4. Kiss A , et al
      . Eligibility criteria of randomized controlled trials published in high-impact general medical journals: a systematic sampling review. JAMA 2007;297:1233–40.doi:10.1001/jama.297.11.1233 pmid:http://www.ncbi.nlm.nih.gov/pubmed/17374817
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      2. Bots SH ,
      3. Siegersma KR ,
      4. Onland-Moret NC , et al
      . Routine clinical care data from thirteen cardiac outpatient clinics: design of the cardiology centers of the Netherlands (CCN) database. BMC Cardiovasc Disord 2021;21:1–9.doi:10.1186/s12872-021-02020-7
      OpenUrlCrossRef
    19. ↵
      2. Diamond GA
      . A clinically relevant classification of chest discomfort. J Am Coll Cardiol 1983;1:574–5.doi:10.1016/S0735-1097(83)80093-X pmid:http://www.ncbi.nlm.nih.gov/pubmed/6826969
      OpenUrlFREE Full Text
    20. ↵
      2. Buuren Svan ,
      3. Groothuis-Oudshoorn K
      . mice : Multivariate Imputation by Chained Equations in R. J Stat Softw 2011;45:1–67.doi:10.18637/jss.v045.i03
      OpenUrlCrossRef
    21. ↵
      2. Austin PC
      . An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res 2011;46:399–424.doi:10.1080/00273171.2011.568786 pmid:http://www.ncbi.nlm.nih.gov/pubmed/21818162
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      2. DE H ,
      3. Imai K ,
      4. King G
      . MatchIt: nonparametric preprocessing for parametric causal inference. J Stat Softw 2011;42:1–28.
      OpenUrlCrossRef
    23. ↵
      2. Conroy RM ,
      3. Pyörälä K ,
      4. Fitzgerald AP , et al
      . Estimation of ten-year risk of fatal cardiovascular disease in Europe: the SCORE project. Eur Heart J 2003;24:987–1003.doi:10.1016/S0195-668X(03)00114-3 pmid:http://www.ncbi.nlm.nih.gov/pubmed/12788299
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      2. Juarez-Orozco LE ,
      3. Saraste A ,
      4. Capodanno D , et al
      . Impact of a decreasing pre-test probability on the performance of diagnostic tests for coronary artery disease. Eur Heart J Cardiovasc Imaging 2019;20:1198–207.doi:10.1093/ehjci/jez054 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30982851
      OpenUrlPubMed
    25. ↵
      2. Poon M ,
      3. Lesser JR ,
      4. Biga C , et al
      . Current evidence and recommendations for coronary cta first in evaluation of stable coronary artery disease. J Am Coll Cardiol 2020;76:1358–62.doi:10.1016/j.jacc.2020.06.078 pmid:http://www.ncbi.nlm.nih.gov/pubmed/32912449
      OpenUrlFREE Full Text
    26. ↵
      2. Lubbers M ,
      3. Dedic A ,
      4. Coenen A , et al
      . Calcium imaging and selective computed tomography angiography in comparison to functional testing for suspected coronary artery disease: the multicentre, randomized crescent trial. Eur Heart J 2016;37:1232–43.doi:10.1093/eurheartj/ehv700 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26746631
      OpenUrlCrossRefPubMed
    27. ↵
      2. Williams MC ,
      3. Hunter A ,
      4. Shah ASV , et al
      . Use of coronary computed tomographic angiography to guide management of patients with coronary disease. J Am Coll Cardiol 2016;67:1759–68.doi:10.1016/j.jacc.2016.02.026 pmid:http://www.ncbi.nlm.nih.gov/pubmed/27081014
      OpenUrlFREE Full Text
    28. ↵
      1. SCOT-HEART investigators
      . CT coronary angiography in patients with suspected angina due to coronary heart disease (SCOT-HEART): an open-label, parallel-group, multicentre trial. Lancet 2015;385:2383–91.doi:10.1016/S0140-6736(15)60291-4 pmid:http://www.ncbi.nlm.nih.gov/pubmed/25788230
      OpenUrlCrossRefPubMed
    29. ↵
      2. Karius P ,
      3. Lembcke A ,
      4. Sokolowski FC , et al
      . Extracardiac findings on coronary computed tomography angiography in patients without significant coronary artery disease. Eur Radiol 2019;29:1714–23.doi:10.1007/s00330-018-5688-4 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30255246
      OpenUrlCrossRefPubMed
    30. ↵
      2. Zeb I ,
      3. Abbas N ,
      4. Nasir K , et al
      . Coronary computed tomography as a cost-effective test strategy for coronary artery disease assessment - a systematic review. Atherosclerosis 2014;234:426–35.doi:10.1016/j.atherosclerosis.2014.02.011 pmid:http://www.ncbi.nlm.nih.gov/pubmed/24769305
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Twitter @InnovatieHester

    • Contributors KRS, NCOM, HMdR and LH designed the study. KRS cleaned the raw data, performed data analyses and wrote the manuscript. NCOM and HMdR supported data analysis and critically reviewed the manuscript. IIT, LH and GAS collected the raw data and reviewed the manuscript. YA, PvdH and JN critically reviewed the manuscript. LH is responsible for the overall contact and acts as the guarantor.

    • Funding This study was funded by the Dutch Heart Foundation (CVON-AI: 2018B017).

    • Competing interests LH, IIT and GAS are employed by Cardiology Centres of the Netherlands.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.