Coronary artery calcium in the population-based ImaLife study
Xia, Congying
DOI:
10.33612/diss.136415357
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Xia, C. (2020). Coronary artery calcium in the population-based ImaLife study: relation to cardiovascular risk factors and cognitive function. University of Groningen. https://doi.org/10.33612/diss.136415357
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
GENERAL DISCUSSION
The coronary artery calcium (CAC) score is widely used to non-invasively quan-tify the amount of coronary atherosclerosis. Many epidemiological studies have shown the strong predictive value of CAC scoring for cardiovascular disease. In this thesis, CAC score distribution in association with cardiovascular risk factors and cognitive function in a general Dutch population was investigated. The clinical relevance of the studies and results described in this thesis have been discussed in the prior chapters. In this chapter, a more general discussion of the main findings is provided. Moreover, methodological considerations are discussed, and potential clinical implications and directions for future research are addressed.
STUDY DESIGN OF THE IMALIFE STUDY
Studies described in this thesis are in the framework of the ImaLife (Imaging in Lifelines) study. In chapter 2, the rationale and study design of the ImaLife study was described. The ImaLife study is an imaging study embedded in the popula-tion-based Lifelines cohort study in the Northern part of the Netherlands. Lifelines is a large multi-generational cohort study as well as a biobank with comprehensive data on lifestyle, environmental, phenotypic, and genomic factors [1]. The baseline enrolment of Lifelines started in 2006. Participants are invited to complete follow-up questionnaires every 1.5 years and to visit the Lifelines research center for exam-inations once every 5 years. Additionally, side studies are performed in subgroups of Lifelines participants. The second-round assessment of Lifelines were completed between 2014 and 2018. After completion of the second-round assessment, par-ticipants aged 45 years and above were invited to participate in the ImaLife study and undergo a low-dose computed tomography (CT) examination of the heart and lungs. The ImaLife study was launched in 2017 and is still ongoing. In total, 12,000 participants will undergo cardiac and chest acquisitions with third-generation du-al-source CT. Subsequently, reference values of imaging biomarkers for the early stages of the big three (Big-3) diseases, namely the CAC score for coronary artery disease (CAD), lung nodule presence and size for lung cancer, and pulmonary density and bronchial wall thickness for chronic obstructive pulmonary disease, can be established for the general Dutch population.
MAIN FINDINGS
Calcium scoring by chest CT
Given the fact that low-dose chest CT is increasingly used in lung cancer screening and CAC is commonly observed on chest scans, there is interest to also perform
cardiovascular risk assessment based on chest CT. This may further increase the benefits of lung cancer screening, since CAC scoring plays an important role in the risk assessment of CAD and risk factors overlap between CAD and lung cancer, in particular smoking [2-7]. In chapter 3, we evaluated the accuracy of CAC scoring based on non-ECG triggered chest CT scanning using third-generation dual-source CT with a high-pitch scan mode, compared to the ECG-triggered cardiac CT scan. We found that low-dose, high-pitch chest CT scanning shows almost perfect agree-ment with the dedicated cardiac CT scan for CAC detection and cardiovascular risk stratification. However, low-dose chest CT underestimated the CAC score in certain cases, while risk categorization in body mass index ≥30 was inaccurate.
Cardiovascular risk factors of CAC
Systematic COronary Risk Evaluation (SCORE), on the basis of classical risk fac-tors, is recommended for cardiovascular risk assessment in Europe [8]. Adding the CAC score to the SCORE evaluation may improve cardiovascular risk prediction. In chapter 4, we examined the association of classical cardiovascular risk factors with the presence of CAC and then compared cardiovascular risk stratification based on SCORE versus CAC scoring. We found that in participants at low risk (SCORE <1%), 32.7% of men and 17.1% of women did have CAC, while in participants at high risk (SCORE ≥5%), 26.9% of men had no CAC (only 0.1% of women had SCORE ≥5%). Moreover, classical cardiovascular risk factors explained only a limited proportion of the presence of CAC.
Given the moderate discriminative ability (assessed by C-statistics) of current risk scores to predict cardiovascular risk [9, 10], it is considered worthwhile to explore novel risk markers that can help improve risk assessment and could also be po-tential interventional targets. Apart from the CAC score, that is now mentioned in guidelines and has a specified role in intermediate risk individuals [4, 8], there are other potentially interesting risk markers. We examined the association between advanced glycation products (AGEs) that are involved in the atherosclerotic pro-cess and CAC (chapter 5). We found that increased skin autofluorescence (SAF), a marker of tissue accumulation of AGEs, was associated with the presence of CAC. The association was largely explained by classical cardiovascular risk factors. In-terestingly, different subpopulations were selected based on high SAF or high CAC score.
CAC and cognitive function
The CAC score reflects the integrated effect of exposure to known and unknown cardiovascular risk factors on the arterial wall, hence the strong association with
cardiovascular diseases. Cardiovascular risk factors also contribute to dementia [11]. Therefore, CAC may also play a role in risk prediction of cognitive impairment and dementia. In this thesis, the association between CAC and cognitive function was investigated. In chapter 6, we reviewed epidemiological studies that investi-gated associations of CAC and clinical CAD with cognitive function and dementia. There was limited evidence from longitudinal cohorts to support that increased subclinical atherosclerosis burden, quantified by the CAC score, is associated with an increased risk of dementia [12-14]. In chapter 7, we investigated the associa-tion between CAC and cognitive funcassocia-tion in a Dutch adult populaassocia-tion. We found an inverse association between CAC level and cognitive performance of working memory in adults aged ≥45 years, independent of cardiovascular risk factors.
METHODOLOGICAL CONSIDERATION
Chest CT versus cardiac CT for calcium scoring
Before calculating CAC scores, visualization of CAC is the first step. ECG-trig-gered cardiac CT is routinely used for CAC scoring, with a standard acquisition and post-processing procedure. Many studies have investigated the feasibility of obtain-ing CAC scores from non-ECG triggered, chest CT, with multiple studies reportobtain-ing consistent results between CAC scores based on chest CT and those based on cardiac CT. However, differences in absolute score were observed in every study that cannot be simply corrected by using calibration factors, in particular underes-timation of the CAC score in chest CT [15-19]. The Society of Thoracic Radiology and the Society of Cardiovascular Computed Tomography advocate that the pres-ence or abspres-ence of CAC based on non-triggered chest CT should be reported, but they do not suggest that non-ECG triggered chest CT can replace ECG-triggered cardiac CT for CAC scoring [20]. Our study (chapter 3) validated CAC scoring based on a high-pitch, non-ECG triggered acquisition with third-generation du-al-source CT. This study showed that risk stratification can be performed based on CAC scores obtained from non-ECG triggered chest CT in non-obese individuals. Besides applying ECG-triggering or not in a CT protocol, also other CT parameter settings can lead to differences in absolute CAC scores. For instance, variation in reconstruction settings such as slice thickness, reconstruction algorithm and kernel can also result in differences in CAC scores [21, 22]. In chapter 3, scans obtained by non-ECG triggered chest CT were reconstructed using the same cardiac calci-um scoring dedicated reconstruction protocol in terms of field-of-view, slice thick-ness and kernel. Chest CT reconstruction protocols normally differ from the cardiac reconstruction; we did not investigate the impact of the routine chest CT recon-struction protocol on CAC score quantification. Therefore, our result is insufficient
to make a recommendation on which reconstruction protocol should be used or whether both are feasible. The final decision awaits a further comparison study.
Cardiovascular risk prediction
Cardiovascular risk prediction is the basis of prevention of CAD. It helps to differ-entiate individuals’ risk levels and to make decisions about initiation of medical treatment. In Europe, SCORE has been developed and recommended for cardio-vascular risk assessment, however, it has a limited discriminative ability [8]. Efforts have been made to explore novel risk markers that can improve risk prediction; CAC is one of the most promising risk modifiers. In this thesis, we first compared the distribution of CAC score categories to SCORE risk categories (chapter 4). More than one-fourth of men who were identified with high risk (based on SCORE ≥5%) could be reclassified to a lower risk level due to absent CAC. While 32.7% of men and 17.1% of women who were at low risk (SCORE <1%) could be reclassi-fied to a higher risk level because of having CAC. This finding suggests that adding the CAC score to SCORE may have a substantial impact on cardiovascular risk classification. Nevertheless, in this study, we were not able to directly evaluate the prognostic value of adding CAC score to SCORE due to lack of long-term follow-up outcomes. To assess the added value of a new risk marker to existing risk scores, statistical measures such as comparison of receiver operating characteristic curves and net reclassification improvement is preferred [23, 24].
The CAC score, an imaging biomarker of subclinical atherosclerosis, can be a surrogate of clinical cardiovascular outcomes. Besides its added value in cardio-vascular risk assessment, it may be helpful to identify new risk factors. We first investigated the association of well-known classical cardiovascular risk factors with the possibility of CAC presence. Population attributable fraction (PAF) was used to estimate the proportion of CAC presence that would be reduced by eliminating exposure to a risk factor. We found that classical risk factors only contributed to a limited proportion of CAC presence (combined PAF was 18.5% in men and 31.4% in women). This finding suggests that besides to refine classical risk factors that are amendable, there is room left for potential new intervention targets to lower cardiovascular risk. We then investigated the relationship between SAF, assessing tissue accumulation of AGEs, and CAC. SAF was associated with CAC, but the as-sociation was largely explained by classical risk factors. There was limited overlap in selected subpopulations with high SAF or high CACS. These findings suggest complementary roles of SAF and CAC scoring in cardiovascular risk assessment, although the added value of SAF to predict cardiovascular events needs to be eval-uated when follow-up outcomes are available.
Pitfalls of cross-sectional analysis
In a cross-sectional study, comparisons of different population groups are made at a single point in time, and measurements of exposure and outcomes ideally are at the same time. This cross-sectional design is comparatively easy to conduct, and can provide information about correlations of different variables with the outcome of interest, which can pave the way for further research. However, the main limitation of the cross-sectional design is that a causal relationship cannot be derived. Given known underlying pathophysiological mechanisms: the fact that AGEs accumula-tion is involved in the atherosclerotic process and the fact that vascular risk factors contribute to dementia [25, 26], it is not unreasonable to assume that SAF, reflect-ing AGEs accumulations, predicts CAC; and that the CAC score, reflectreflect-ing cumu-lative effects of exposure to cardiovascular risk factors, predicts cognitive impair-ment. However, our cross-sectional results are insufficient to prove that increased AGEs accumulation directly causes subclinical atherosclerosis, nor that subclinical atherosclerosis is a direct cause of cognitive impairment. Future longitudinal cohort studies are needed to prove causality and etiology.
Furthermore, estimated associations in a cross-sectional study may be unintention-ally distorted when measures of outcome and exposure are not taken at the same time, and outcome or exposure could have changed in between. In the studies of this thesis a cross-sectional design was used to evaluate CAC in relation to clas-sical cardiovascular risk factors, SAF and cognitive function. The magnitude of associations as reported might be biased due to the time lag between CAC scoring and measurements of either classical cardiovascular risk factors, SAF or cognitive function. More specifically, the CAC scores were obtained up to 3 years after as-sessment of classical cardiovascular risk factors and 4 to 10 years after measure-ment of SAF. CAC can develop during the time lag, it could be the case that the CAC was absent if it had been assessed at the same time point as the measure-ments of cardiovascular risk factors and SAF. Therefore, associations of classical cardiovascular risk factors and SAF with CAC presence may be overestimated in our studies. Besides, measurements of cognitive function were 2 to 6 years before CAC scoring. The association between CAC and cognitive function may be under-estimated in our study, since participants with worse cognitive impairment may fail to participate in the CAC score assessment.
Generalizability
Generalizability in epidemiology is defined as “unbiased inferences regarding a tar-get population” [27]. The study sample is a key component that needs to be taken into account when making inferences based on the results of a given study. The
study sample should be representative of the whole population that is being studied [28]. However, it is questionable how to define a population. The ImaLife study is an imaging study embedded in the Lifelines, which is a large population-based cohort comprising of inhabitants of the northern part of the Netherlands and their families [1]. In this case, the population was defined by establishing its boundary, mainly the regional boundary: the Northern part of the Netherlands. It is of note that nearly all of this population is Caucasian. Therefore, caution is needed when generalizing described results to different regions and ethnics. Furthermore, in the ImaLife study, age is another boundary. In the ImaLife study given that early stages of Big-3 diseases are rare below 45 years [29-31], only adults aged 45 years and older were recruited. In chapter 4, the magnitude of associations between classical cardiovascular risk factors and the presence of CAC was evaluated in a middle-aged (aged 45-60 years) Dutch population. The results cannot be directly generalized to a population outside this age range.
Besides the study sample, also the measurement or determination of the outcome or exposure should be considered before generalizing the results. In chapter 3, we investigated the accuracy of CAC scoring by using a non-ECG triggered chest scan based on the third-generation, dual-source CT system with the high-pitch mode. The result of this study cannot be generalized to other CT systems, especially to those that cannot achieve this high-pitch mode.
CLINICAL IMPLICATIONS AND FUTURE DIRECTIONS One scan for the Big-3 diseases
Shared risk factors among the Big-3 diseases and visualization of the lungs and heart on one chest CT scan inspired the idea of combining the early detection for the Big-3 diseases in one low-dose CT examination. Imaging biomarkers for early stages of CAD (CAC score), lung cancer (presence and size of lung nodules), and chronic obstructive pulmonary disease (pulmonary density and bronchial wall thickness) have been proposed. Currently, imaging biomarkers for the Big-3 dis-eases are obtained from separate scans that are acquired and reconstructed with optimized protocols for each imaging biomarker. Combining the detection of the Big-3 disease within on chest scan acquisition can reduce radiation dose and may increase effectiveness. Further studies can be focused on integrating a 3-in-1 ac-quisition protocol for imaging biomarkers of the Big-3 diseases that can be applied to all participants.
CAC in primary prevention of CAD
Dutch population is presented (chapter 4). This paves the way for setting CAC score reference values, which is a prerequisite for using CAC scoring in primary prevention of CAD. Reference values are used to differentiate individuals’ risk levels and cutoffs of positive tests to direct downstream work-up. Moreover, we found only a limited overlap in risk categorization between classical risk factors based SCORE and CAC scoring. This result suggests that adding CAC scoring to SCORE could have a considerable impact on cardiovascular risk stratification. Population-based studies have shown that the addition of CAC score to classical risk factors had the largest improvements in discrimination and risk classification; a CAC score of zero resulted in the largest reduction of post-test risk and resulted in the most accurate downward risk reclassification compared to other negative cardiovascular risk markers [32, 33]. The long-term follow-up in the ImaLife study have to be awaited before the added value of CAC scores for risk prediction in the adult Dutch population can be determined. In addition, to answer the questions of whether CAC scoring can lead to a reduction in cardiovascular events and mortal-ity, and of whether CAC score screening is cost-effective, results from prospective randomized control trials are needed (like the ROBINSCA trial).
Novel risk markers for CAD
In chapter 5, the relationship between SAF, measuring tissue accumulation of AGEs, and CAC was assessed. Our cross-sectional results showed that increased SAF was associated with increased probability of CAC presence, but this associ-ation was largely explained by classical cardiovascular risk factors. At the same time, we found only limited overlap in subgroups with high SAF or high CACS. These findings suggest that SAF and CAC may reflect the heterogeneity of the ath-erosclerotic process, and have complementary roles in identifying individuals at el-evated cardiovascular risk. Further longitudinal studies can compare the predictive value of SAF versus CAC scores for cardiovascular events and to investigate the added value of SAF and CAC over classical risk factors. Furthermore, CAC score is a surrogate for cardiovascular events, with a continuous instead of dichotomous outcome. The Lifelines study in which ImaLife is embedded, collects comprehen-sive data on environmental, genomic, and phenotypical factors. Therefore, future studies can aim to evaluate novel risk factors, such as genetic markers, in relation to CAC as subclinical cardiovascular outcome.
Coronary artery calcium and cognitive function
In chapter 7, our results revealed an inverse association between CAC level and cognitive function. This suggests that CAC scoring may help to identify individuals at risk of cognitive impairment. Prevention or halting of subclinical atherosclerosis
may help to delay or prevent dementia. However, further experimental studies are needed to illustrate the molecular mechanisms linking atherosclerosis and demen-tia to be able to determine new potendemen-tial targets for intervention. Moreover, in this thesis, the cognitive function was measured by a collection of computerized tests whereas the brain structure was not assessed. In the future, a subgroup of ImaL-ife participants will be invited to participate in the MemolImaL-ife study to undergo brain magnetic resonance imaging. The Memolife study will enable a better understand-ing of how CAC is associated with pathological structural changes in the brains and consequently impact the cognitive function.
CONCLUSION
To conclude, we described the study design of an imaging study that is aimed to establish population-based reference values of CAC scores as well as imaging bio-markers for early stages of lung cancer and chronic obstructive pulmonary disease in this thesis. With the third-generation dual-source high-pitch technique, one chest CT acquisition for both lung and cardiac CT screening is feasible in non-obese in-dividuals. This option of CT acquisition reduces the radiation dose. SCORE, based on cardiovascular risk factors, was found to be related to the CAC score, but only a limited proportion of CAC in the middle-aged population would be prevented if exposure to classical risk factors was eliminated. Adding CAC scoring to SCORE could have a considerable impact on cardiovascular risk stratification. AGEs accu-mulation is associated with the presence of CAC, but largely explained by classical risk factors; AGEs and CAC seem complementary in identifying high-risk subpop-ulations. Last but not the least, CAC score is promising in prediction of cognitive impairment although a longitudinal study is needed to confirm this.
REFERENCES:
1 Scholtens S, Smidt N, Swertz MA, et al. Cohort Profile: LifeLines, a three-generation cohort study and bio-bank. International journal of epidemiolo-gy 2015;44:1172-80.
2 Hecht HS, Henschke C, Yan-kelevitz D, et al. Combined detection of coronary artery disease and lung cancer. European heart journal 2014;35:2792-6. 3 Polonsky TS, McClelland RL, Jorgensen NW, et al. Coronary artery calcium score and risk classification for coronary heart disease prediction. Jama 2010;303:1610-6.
4 Bittner VA. The New 2019 ACC/ AHA Guideline on the Primary Preven-tion of Cardiovascular Disease [pub-lished online ahead of print, 2019 Mar 17] Circulation 2019.
5 Vliegenthart R, Oudkerk M, Hofman A, et al. Coronary calcification improves cardiovascular risk prediction in the elderly. Circulation 2005;112:572-7.
6 Detrano R, Guerci AD, Carr JJ, et al. Coronary calcium as a predictor of coronary events in four racial or ethnic groups. The New England journal of medicine 2008;358:1336-45.
7 McClelland RL, Jorgensen NW, Budoff M, et al. 10-Year Coronary Heart Disease Risk Prediction Using Coronary Artery Calcium and Traditional Risk Fac-tors: Derivation in the MESA (Multi-Eth-nic Study of Atherosclerosis) With Validation in the HNR (Heinz Nixdorf Recall) Study and the DHS (Dallas Heart
Study). Journal of the American College of Cardiology 2015;66:1643-53.
8 Piepoli MF, Hoes AW, Agewall S, et al. 2016 European Guidelines on car-diovascular disease prevention in clinical practice: The Sixth Joint Task Force of the European Society of Cardiology and Other Societies on Cardiovascular Disease Prevention in Clinical Practice (constituted by representatives of 10 societies and by invited experts)Devel-oped with the special contribution of the European Association for Cardiovascular Prevention & Rehabilitation (EACPR). European heart journal 2016;37:2315-81.
9 Pletcher MJ, Sibley CT, Pignone M, et al. Interpretation of the coronary artery calcium score in combination with conventional cardiovascular risk factors: the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation 2013;128:1076-84. 10 Rana JS, Gransar H, Wong ND, et al. Comparative value of coronary ar-tery calcium and multiple blood biomark-ers for prognostication of cardiovascular events. The American journal of cardiol-ogy 2012;109:1449-53.
11 Qiu C, Fratiglioni L. A major role for cardiovascular burden in age-related cognitive decline. Nature reviews Cardi-ology 2015;12:267-77.
12 Bos D, Vernooij MW, de Bruijn RF, et al. Atherosclerotic calcification is related to a higher risk of dementia and cognitive decline. Alzheimer’s & dementia : the journal of the Alzheimer’s
Association 2015;11:639-47.e1. 13 Fujiyoshi A, Jacobs DR, Jr., Fitzpatrick AL, et al. Coronary Artery Calcium and Risk of Dementia in MESA (Multi-Ethnic Study of Atherosclerosis). Circulation Cardiovascular imaging 2017;10.
14 Kuller LH, Lopez OL, Mackey RH, et al. Subclinical Cardiovascular Disease and Death, Dementia, and Coronary Heart Disease in Patients 80+ Years. Journal of the American College of Cardiology 2016;67:1013-22.
15 Chen Y, Hu Z, Li M, et al. Comparison of Nongated Chest CT and Dedicated Calcium Scoring CT for Coronary Calcium Quantification Using a 256-Dector Row CT Scanner. Academic radiology 2019.
16 Wu M-T, Yang P, Huang Y-L, et al. Coronary Arterial Calcification on Low-Dose Ungated MDCT for Lung Can-cer Screening: Concordance Study with Dedicated Cardiac CT. American Journal of Roentgenology 2008;190:923-8. 17 Budoff MJ, Nasir K, Kinney GL, et al. Coronary artery and thoracic cal-cium on noncontrast thoracic CT scans: Comparison of ungated and gated examinations in patients from the COPD Gene cohort. Journal of cardiovascular computed tomography 2011;5:113-8. 18 Xie X, Zhao Y, De Bock GH, et al. Validation and prognosis of coro-nary artery calcium scoring in nontrig-gered thoracic computed tomography: Systematic review and meta-analysis. Circulation: Cardiovascular Imaging 2013;6:514-21.
19 Hutt A, Duhamel A, Deken V, et al. Coronary calcium screening with dual-source CT: reliability of ungated, high-pitch chest CT in comparison with dedicated calcium-scoring CT. European radiology 2016;26:1521-8.
20 Hecht HS, Cronin P, Blaha MJ, et al. 2016 SCCT/STR guidelines for coronary artery calcium scoring of noncontrast noncardiac chest CT scans: A report of the Society of Cardiovascu-lar Computed Tomography and Soci-ety of Thoracic Radiology. Journal of cardiovascular computed tomography 2017;11:74-84.
21 Vonder M, van der Werf NR, Leiner T, et al. The impact of dose re-duction on the quantification of coronary artery calcifications and risk categori-zation: A systematic review. Journal of cardiovascular computed tomography 2018;12:352-63.
22 Vliegenthart R, Song B, Hofman A, et al. Coronary calcification at elec-tron-beam CT: effect of section thickness on calcium scoring in vitro and in vivo. Radiology 2003;229:520-5.
23 Pencina MJ, D’Agostino RB, Sr., D’Agostino RB, Jr., et al. Evaluating the added predictive ability of a new mark-er: from area under the ROC curve to reclassification and beyond. Statistics in medicine 2008;27:157-72; discussion 207-12.
24 Cook NR, Ridker PM. Advanc-es in measuring the effect of individual predictors of cardiovascular risk: the role of reclassification measures. Annals of internal medicine 2009;150:795-802.
25 Attems J, Jellinger KA. The overlap between vascular disease and Alzheimer’s disease--lessons from pa-thology. BMC medicine 2014;12:206. 26 Singh R, Barden A, Mori T, et al. Advanced glycation end-products: a review. Diabetologia 2001;44:129-46. 27 Ponsonby AL, Dwyer T, Couper D. Is this finding relevant? Generalisa-tion and epidemiology. Aust N Z J Public Health 1996;20:54-6.
28 Kukull WA, Ganguli M. Gen-eralizability: the trees, the forest, and the low-hanging fruit. Neurology 2012;78:1886-91.
29 Hoff JA, Chomka EV, Krainik AJ, et al. Age and gender distributions of coronary artery calcium detected by electron beam tomography in 35,246 adults. The American journal of cardiolo-gy 2001;87:1335-9.
30 Lung cancer figures and context current situation public healthcare.info. 31 Boudewijn IM, Postma DS, Telenga ED, et al. Effects of ageing and smoking on pulmonary computed tomog-raphy scans using parametric response mapping. The European respiratory journal 2015;46:1193-6.
32 Lin JS, Evans CV, Johnson E, et al. Nontraditional Risk Factors in Cardio-vascular Disease Risk Assessment: Up-dated Evidence Report and Systematic Review for the US Preventive Services Task Force. Jama 2018;320:281-97. 33 Blaha MJ, Cainzos-Achirica M, Greenland P, et al. Role of Coronary Artery Calcium Score of Zero and Other Negative Risk Markers for
Cardiovas-cular Disease: The Multi-Ethnic Study of Atherosclerosis (MESA). Circulation 2016;133:849-58.
