Smokers suffer from a substantial increase in both cardiovascular and malignancy-related mortality. In 2016, the United States Preventive Services Task Force (USPSTF) recommended that current or past smokers aged 55-80 years at risk of lung cancer who had a smoking history of at least 30 pack-years would undergo annual low-dose computed tomography (CT) for lung cancer screening. In this subgroup of smokers, atherosclerotic cardiovascular disease (ASCVD) risk is high. However, methods of ASCVD risk prediction in this subgroup are less studied and the accuracy of common tools of risk prediction such as the pooled cohort risk equations (PCE) is less clear. This is important in the decision regarding statin and aspirin treatment.

In addition to reporting data useful for lung cancer detection, low-dose CT scanning can estimate coronary artery calcium (CAC), a marker of ASCVD. Although prior reports demonstrated an association between CAC obtained by low-dose CT and ASCVD events in all-smokers and lung cancer screening eligible (LCSE) subjects, prior studies had limited reporting of ASCVD risk factors. This in turn prevented comparing CAC with other commonly used risk predictors such as the PCE.

In a report by Leigh et al.,¹ the authors used the Multi-Ethnic Study of Atherosclerosis (MESA) population, a well-characterized prospective cohort, to examine the usefulness of PCE and CAC as risk predictors of ASCVD events in all-smokers and LCSE subjects. They also address the ability of CAC to improve PCE in predicting ASCVD events in these subgroups.

The authors analyzed data from 3,356 MESA participants. Those with history of myocardial infarction, angina, heart failure, prior stroke or transient ischemic attack or prior invasive vascular procedures were excluded. All-smokers were defined as current or past smokers according to the standard American Thoracic Society questionnaire. LCSE subjects were identified according to the USPSTF criteria (described above).

In MESA, CAC was measured using non-contrast gated cardiac CT with either electron beam CT or a multidetector CT. A composite primary outcome consisted of adjudicated ASCVD events of fatal and non-fatal myocardial infarction, other fatal and non-fatal coronary heart disease, fatal and non-fatal cerebrovascular disease.

After a follow up of 11 years, the majority of the events were fatal and non-fatal coronary heart disease. Both LCSE subjects and all-smokers experienced high rates of ASCVD events with LCSE subjects having higher risk of ASCVD events (21%) compared to all-smokers (13%). The LCSE subgroup still had higher risk of ASCVD events compared to all-smokers in subjects with no CAC (14% vs. 7%). This suggests that LCSE subjects are a higher risk subgroup requiring aggressive risk optimization.

An important finding in the study is low rate of statin prescription in the LCSE subgroup. While 82.5% of LCSE subjects were statin eligible, only 19% were on statin at baseline indicating a significant gap in care which requires further attention and improvement.

Although both PCE and CAC were independent predictors of ASCVD events in all-smokers and the LCSE subgroup, both PCE and CAC were stronger predictors of ASCVD risk in all-smokers than in LCSE subjects. Perhaps one of the most interesting findings in this study is evidence that the absence of CAC is not an indicator of low overall ASCVD risk in the LSCE population. The authors show that even when CAC is absent, 7% of all-smokers and 14% of LCSE subjects have ASCVD events. Furthermore, in LCSE subjects eligible for statins who had CAC = 0, 21% had ASCVD events.

When predicting risk of ASCVD events, PCE had a higher discriminative ability in all-smokers compared to LCSE subjects. Both all-smokers and LCSE subjects had a high percentage of CAC >0 indicating high prevalence of subclinical atherosclerosis. In addition, when comparing LCSE subject to all-smokers in the same PCE risk category, those in the LCSE subgroup had a higher percentage of CAC >0 compared to all-smokers. Thus, while PCE includes smoking status as a predictor of ASCVD risk, the absence of duration or frequency of smoking in PCE may explain the differences in CAC and ASCVD events in all-smokers compared to LCSE subjects in the same PCE risk category.

The authors used two statistical methods to assess the ability of CAC to improve ASCVD risk prediction beyond PCE. First, they used the receiver operator characteristic (ROC) curve analysis, which showed that the addition of CAC to PCE improved discrimination for ASCVD events in all-smokers but did poorly in LCSE subjects. Second, the authors performed a net reclassification improvement analysis to assess the potential of CAC added to PCE to improve discrimination for ASCVD events in individuals classified low, intermediate or high risk.

In both all-smokers and LCSE subjects with ASCVD events, CAC added to PCE appropriately up classified subjects into a higher risk category compared to using PCE alone. However, in all-smokers and LCSE subjects with no ASCVD events, adding CAC to PCE led to inappropriate re-classification of subjects into a higher risk group compared to PCE alone. Therefore, in all-smokers and the LCSE groups, the addition of CAC to PCE resulted in improved risk reclassification only in those with future events. The authors concluded that in all-smokers and LCSE subjects, adding CAC to PCE did not improve reclassification of risk categories beyond PCE.

Since the CT methods of obtaining CAC in MESA are different from low-dose CT used for lung cancer screening, several differences arise as a result of lack of electrocardiogram gating and lower radiation dosages used in low-dose CT. First, mean CAC values reported by non-gated CT scans are generally lower than CAC reported by gated CT scans.² Also, when CAC is very low, non-gated CT scans are less likely to differentiate between CAC = 0 and CAC >0.^2,3

Although non-gated low-dose CT is shown to accurately estimate CAC category (e.g., CAC <300 vs. >300), non-gated CT is more likely to shift subjects into a lower risk category.^3,4 Due to these limitations, large, well-characterized prospective registries of those undergoing lung cancer screening are needed to confirm the findings of this study using data from low-dose CT scans.

In conclusion, LCSE subjects are a high-risk group in whom PCE and CAC may have significant limitations for ASCVD risk predictions. Furthermore, in this group, CAC did not improve PCE's discriminative ability for ASCVD. CAC should be interpreted with caution in statin eligible LCSE subjects as the absence of CAC is not associated with low overall ASCVD risk in this subgroup. Given the limitations in ASCVD risk prediction with PCE and CAC in the LCSE subgroup and the need for aggressive ASCVD risk optimization, the authors conclude that statins should be considered in those eligible for lung cancer screening regardless of their calculated 10-year ASCVD risk or CAC score.

References

1. Leigh A, Mcevoy JW, Garg P, et al. Coronary artery calcium scores and atherosclerotic cardiovascular disease risk stratification in smokers: MESA. JACC Cardiovasc Imaging 2018. [Epub ahead of print]
2. Arcadi T, Maffei E, Sverzellati N, et al. Coronary artery calcium score on low-dose computed tomography for lung cancer screening. World J Radiol 2014;6:381-7.
3. Hecht HS. Coronary artery calcium analysis and reporting on noncontrast chest CT scans: a paradigm shift in prevention. Curr Cardiovasc Imaging Rep 2016;9:11.
4. Chiles C, Duan F, Gladish GW, et al. Association of coronary artery calcification and mortality in the National Lung Screening Trial: a comparison of three scoring methods. Radiology 2015;276:82–90.

Keywords: Dyslipidemias, Ischemic Attack, Transient, Hydroxymethylglutaryl-CoA Reductase Inhibitors, ROC Curve, Risk Factors, Coronary Vessels, Early Detection of Cancer, Lung Neoplasms, Atherosclerosis, Stroke, Coronary Disease, Angina Pectoris, Myocardial Infarction, Heart Failure, Electrocardiography, Tomography, X-Ray Computed, Smoking, Registries, Radiation Dosage, Tomography

< Back to Listings