key: cord-0936928-6t2yoho0
authors: Gupta, Yogesh Sean; Finkelstein, Mark; Manna, Sayan; Toussie, Danielle; Bernheim, Adam; Little, Brent P.; Concepcion, Jose; Maron, Samuel Z.; Jacobi, Adam; Chung, Michael; Kukar, Nina; Voutsinas, Nicholas; Cedillo, Mario A.; Fernandes, Ajit; Eber, Corey; Fayad, Zahi A.; Hota, Partha
title: Coronary artery calcification in COVID-19 patients: an imaging biomarker for adverse clinical outcomes
date: 2021-02-11
journal: Clin Imaging
DOI: 10.1016/j.clinimag.2021.02.016
sha: 99fed3b8821473f248fb4f03be825571ac1d99d4
doc_id: 936928
cord_uid: 6t2yoho0

BACKGROUND: Recent studies have demonstrated a complex interplay between comorbid cardiovascular disease, COVID-19 pathophysiology, and poor clinical outcomes. Coronary artery calcification (CAC) may therefore aid in risk stratification of COVID-19 patients. METHODS: Non-contrast chest CT studies on 180 COVID-19 patients ≥ age 21 admitted from March 1, 2020 to April 27, 2020 were retrospectively reviewed by two radiologists to determine CAC scores. Following feature selection, multivariable logistic regression was utilized to evaluate the relationship between CAC scores and patient outcomes. RESULTS: The presence of any identified CAC was associated with intubation (AOR: 3.6, CI: 1.4–9.6) and mortality (AOR: 3.2, CI: 1.4–7.9). Severe CAC was independently associated with intubation (AOR: 4.0, CI: 1.3–13) and mortality (AOR: 5.1, CI: 1.9–15). A greater CAC score (UOR: 1.2, CI: 1.02–1.3) and number of vessels with calcium (UOR: 1.3, CI: 1.02–1.6) was associated with mortality. Visualized coronary stent or coronary artery bypass graft surgery (CABG) had no statistically significant association with intubation (AOR: 1.9, CI: 0.4–7.7) or death (AOR: 3.4, CI: 1.0–12). CONCLUSION: COVID-19 patients with any CAC were more likely to require intubation and die than those without CAC. Increasing CAC and number of affected arteries was associated with mortality. Severe CAC was associated with higher intubation risk. Prior CABG or stenting had no association with elevated intubation or death.

Using an institutional COVID-19 dataset, 198 admitted patients ≥ the age 21 with NCCT within a 3-month period prior to and after admissions were identified. Patients with prior coronary artery bypass graft (CABG) surgery or coronary arterial stenting, determined by radiologist suspicion with electronic medical record (EMR) confirmation, were excluded from CAC scoring but were included in outcomes analysis. NCCTs on which at least one radiologist suspected coronary stenting that could not be confirmed or repudiated by EMR were excluded (n = 8).

NCCTs deemed too limited by motion artifact for adequate CAC assessment by at least one radiologist were excluded (n = 7). After exclusions, 180 patients were included for outcomes analysis ( Figure 1) .

Patient information was collected from the EMR. Demographic variables included age, gender, self-reported race and ethnicity. Clinical variables included past medical history (as shown in Table 1 ), body mass index (BMI), smoking history, echocardiography results, laboratory variables including troponins, and selected outpatient medications (statins, antiplatelet regimen, anticoagulant regimen). Antithrombotic therapy was classified as single antiplatelet (SAPT), dual antiplatelet (DAPT), or oral anticoagulant (OAC) [14] . The statin regimen was categorized as combined low/moderate or high intensity following available ACC guidelines [16] [17] [18] . The highest available D-dimer level during admission was recorded (D-dimer level > 1 μg/mL was considered elevated) [19, 20] . Echocardiography results were obtained from EMR. End results of every admission (discharge or death) were recorded for every encounter.

J o u r n a l P r e -p r o o f All NCCTs contained 1-mm-thick slices. They were obtained from five sites in New York City.

In Brooklyn, 108 patients were scanned on Aquillon PRIME (Toshiba, Tokyo, Japan). In Manhattan, 39 patients were scanned on Somatom Definition AS+ (Siemens, Munich, Germany), 14 patients were scanned on iCT 256 (Philips, Amsterdam, The Netherlands), and 4 patients were scanned on Revolution EVO (General Electric, Chicago, USA). In Queens, 15 patients were scanned on Revolution HD (General Electric, Chicago, USA).

Two fellowship-trained cardiothoracic radiologists (Y.S.G. and J.C.; total experience of 12 years) with extensive experience in CAC assessment independently scored each NCCT on soft tissue window of a standard PACS workstation, blinded to patient histories other than COVID-19 positivity. Visual assessment of CAC was performed according to an established ordinal scoring method [21] . Each of the four main coronary arteries was identified (left main, left anterior descending, left circumflex, and right coronary arteries). Calcium extent was scored as 0, 1, 2, or 3 ( Figure 2 ). A score of 1 is defined as involvement of less than one third of the vessel length; 2 as involvement of one third to two thirds of the vessel length; 3 as greater than two thirds of the vessel length. These scores were summed to obtain a total ordinal score of 0-12 for each scan. The total scores were categorized as absent CAC for score 0, mild for score 1-3, moderate for a score 4-5, and severe for score ≥ 6.

When patients had changes indicative of prior coronary artery bypass graft (CABG) surgery or stent placement, confirmed on EMR, CAC scoring was not performed.

J o u r n a l P r e -p r o o f Weighted Cohen's kappa coefficient was used to assess radiologists' agreement in scoring. For individual vessels, stent and CABG cases were excluded in this calculation, due to these being unscorable. The mean of the total scores was calculated and categorized based on above described parameters. In order to ensure usability of as many records as possible in multivariable analysis, missing BMI (11/180, 6.1%) were imputed using predictive mean matching using models that included the outcomes of interest (intubation status, death), demographic information (age, ethnicity, and race), clinical variables (smoking status and comorbidities listed in Table 1) , and the CAC category. These values were then utilized in the multivariable model through multiple imputation according to Rubin's rules [22] . Bivariate analysis of continuous variables (BMI and age) was performed using the Kruskal-Wallis H Test.

Bivariate analysis of categorical variables including race, gender, smoking history, and comorbidities was performed using chi-squared test. These comparisons were performed in order to ascertain statistically significant differences in relation to the primary outcomes of interest (intubation and death). Univariable logistic regression was utilized to estimate the relative effect of variables by calculating unadjusted odds ratios (UOR) for categorical covariates in relation to the outcomes of interest. Random forest was used for feature selection utilizing the Boruta approach [23] . This approach was chosen because it is an all-relevant variable selection method as opposed to a minimal optimal method while providing valuable information concerning the relative importance of features. Feature selection was applied separately to each individual adverse outcome. A multivariable logistic regression model adjusting for selected features was performed for the outcomes of intubation, death, and elevated D-dimer. Two such models were created, one involving the categorical variable of CAC category and one involving the binary variable of presence of any calcification. Adjusted odds ratios (AOR) are presented from these J o u r n a l P r e -p r o o f Journal Pre-proof models. A p-value of less than 0.05 (two-tailed) was considered statistically significant. All analysis was completed using R version 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria).

Population characteristics, comorbidities, and medications: were on an antithrombotic regimen. Of the patients with CAD, 9/25 (36%) were found to have stents and/or CABG. Admitted patient CAC scores, echocardiography findings, and laboratory parameters in relation to intubation and expiration are shown in Table 2 . Additional data relating lab and echocardiography findings in relation to the outcomes of interest are provided in Supplemental Table 1 . 

Fifty-four patients (30%) were intubated during their admission, 59 (33%) died, and 95 (86%) had elevated D-dimer (as defined by > 1 µg/L). Results of feature selection for the multivariable analysis involving CAC categories showing the relative calculated importance of different features is displayed in Figures 3 and 4 . Every category of CAC score was associated with increased odds of intubation in reference to absence of CAC (mild -AOR: 3.3, CI: 1.1-9.8, moderate -AOR: 6.5, CI 1.9-24, severe -AOR 4.0, CI 1. [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] . A severe CAC score was associated with increased odds of mortality (AOR: 5.1, CI: 1.9-15). Additionally, in the model utilizing CAC category, morbid obesity was associated with increased risk of both intubation (AOR: 3.3, CI: 1.1-10) and mortality (AOR: 5.4, CI: 1. [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] . No statistically significant association was found between the category of patients with stent and/or CABG and the outcomes of intubation (AOR: 1.9, CI: 0.4-7.7) or mortality (AOR: 3.4, CI: 1.0-12). For complete univariable and multivariable logistic regres sion analysis results, see Table 3 .

Increasing CAC score was found to be associated with mortality (UOR: 1. There was no association between statin regimen and antithrombotic regimen and the outcomes of interest (Table 3) .

Both mild/moderate (AOR: 5.7, CI: 1.1-40) and severe (AOR: 11, CI: 1.4-124) CAC score categories were found to be associated with an elevated D-dimer. No association was found between CAC categories and score in relation to acute kidney injury (AKI), elevated troponin, and elevated liver function tests. Patients on higher intensity statins had a higher CAC score 

In this retrospective analysis, we studied the relationship between the presence and extent of CAC with adverse outcomes in COVID-19 patients. Our results indicate that COVID-19 patients with any degree of CAC were more likely to require intubation or expire than those without CAC. Higher CAC score was associated with increased mortality. Patients with severe CAC as opposed to patients with lesser degrees of CAC had a higher risk of intubation compared to patients without CAC. These findings are concordant with a study by Dillinger et al., in which the presence and extent of CAC was associated with worse prognosis in hospitalized patients with COVID-19 [24] .

In our cohort, the presence and severity of CAC as independent predictors of mortality and intubation were superior to several reported risk factors including sex, diabetes, hypertension, atrial fibrillation, heart failure, and chronic obstructive pulmonary disease [20, 25, 26] . Number of vessels with CAC and visual score was associated with poor outcomes. This association might reflect the effects of a severe inflammatory cascade resulting in destabilization of pre-existing cardiovascular disease and subsequent mixed shock, as suggested by recent literature [27, 28] . It may also point to underlying chronic inflammation that predisposes to atherosclerotic disease and to acute proinflammatory states such as the cytokine storm.

Although studies have found that COVID-19 patients with CAD are more likely to have myocardial injury, we did not find a similar association in patients with a pre-admission diagnosis of CAD (25/180), which may be attributable to under-diagnosis given the large group of patients with visually moderate or severe CAC (65/180)[1,29]. We did not find a significant association between CAC and troponin elevation, although this may be secondary to relatively small sample size. The association between CAC and mortality in our study in the absence of clinical evidence of myocardial injury suggests the predictive value of CAC may not be limited to the cardiac domain, reflecting host propensity for widespread proinflammatory response and endothelial dysfunction. In autopsy studies of COVID-19 patients, endothelial injury, perivascular inflammation, and/or microthrombosis have recently been implicated as dominant findings in multiple organs, including the lungs, brain, kidneys, skin, and gastrointestinal tract [30] [31] [32] .

We found no statistically significant association between CABG or stent placement and intubation or death in our cohort. Recent literature suggests that hospitalized COVID-19 patients J o u r n a l P r e -p r o o f on anticoagulation have reduced mortality risk while those on statins may benefit [29, 33] .We hypothesized that the CABG/stent patients may have benefitted from such medications as they are more likely to be on them. However, upon analysis of our entire cohort, we found these therapies did not confer an advantage. One possibility is that CABG/stented patients benefited from surgical/procedural management of obstructive CAD in this regard. In keeping with prior studies, our data demonstrates a positive association of intensity of statin regimen with CAC scores [34, 35] .

Our results indicate a significant role for assessing CAC severity and extent in COVID- NCCTs are sometimes performed on hospitalized COVID-19 patients for various reasons; therefore, visual CAC scoring does not require additional imaging assessment but can provide useful prognostic data [36] . Although CAC is an established imaging marker for predicting cardiovascular event risk, CAC reporting incidence may be as low as 1% [36, 37] .

CAC on NCCT may be the first evidence of undiagnosed CAD and may be a more objective and dependable means of rapid risk stratification in COVID-19 patients, unaffected by confounding variables such as recall bias that may occur during history taking. We recommend radiologists report CAC in all COVID-19 patients to best guide clinical management.

Limitations of this study include its retrospective nature, which may introduce observer bias in CAC assessment. Our cohort was limited to patients who had a NCCT, which may impart selection bias and may not represent the full array of associations of CAC with COVID-19 outcomes. Furthermore, as described previously, patients with NCCT within a 3-month period prior to and after admission were included, which may have resulted in higher relative inclusion of patients who survived. Future investigation may incorporate a cohort with more diverse imaging to determine an association of CAC with other adverse events, including VTE, stroke, and bowel ischemia.

Visual CAC scoring on routine NCCTs in hospitalized COVID-19 patients is a rapid and reproducible imaging biomarker that may provide an independent assessment of risk of intubation and death, with those having severe CAC at higher risk. Patients treated with CABG/stents did not incur a similar risk, possibly reflecting a combination of treated CAD and medical therapy. Assessing CAC severity on NCCTs imparts valuable information regarding risk prognostication and may guide management in COVID-19 patients. J o u r n a l P r e -p r o o f This diagram demonstrates how we arrived at the total number of patients included in the outcomes analysis for our retrospective cohort study. 

Boxplot showing distribution of attribute importance relative to the outcome of intubation over runs. Black and white boxplots represent Z scores of rejected and confirmed attributes respectively. Confirmed attributes were included in multivariable analysis. CAD=coronary artery disease; COPD=chronic obstructive pulmonary disease; CKD=chronic kidney disease; CAC=coronary artery calcification.

Boxplot showing distribution of attribute importance relative to the outcome of death over runs. Black and white boxplots represent Z scores of rejected and confirmed attributes respectively. Confirmed and tentative attributes were included in multivariable analysis. CAD=coronary artery disease; COPD=chronic obstructive pulmonary disease; CKD=chronic kidney disease; CAC=coronary artery calcification.

J o u r n a l P r e -p r o o f Journal Pre-proof Highlights: 1. The presence of any coronary artery calcifications on Chest CT imaging were found to be associated with increased odds of intubation and mortality in admitted patients with confirmed COVID-19. 2. Increased degree of coronary artery calcifications and increasing number of involved vessels was associated with increased odds of mortality. 3. Coronary artery calcification scoring on non-ECG-gated noncontrast CT chest studies was feasible in almost all COVID-19 patients and demonstrated high inter-reader agreement. 

Prevalence and Impact of Myocardial Injury in Patients Hospitalized with COVID-19 Infection n

Prevalence and impact of cardiovascular metabolic diseases on COVID-19 in China

Association of Cardiac Injury With Mortality in Hospitalized Patients With COVID-19 in Wuhan, China

Cardiac injury is associated with mortality and critically ill pneumonia in COVID-19: A meta-analysis

Evaluation of novel coronavirus disease (COVID-19) using quantitative lung CT and clinical data: prediction of short-term outcome

ACCF/AHA 2007 clinical expert consensus document on coronary artery calcium scoring by computed tomography in global cardiovascular risk assessment and in evaluation of patients with chest pain: a report of the American College of Cardiology Foundation Clinical Expert Consensus Task Force (ACCF/AHA Writing Committee to Update the 2000 Expert Consensus Document on Electron Beam Computed Tomography) developed in collaboration with the Society of Atherosclerosis Imaging and Prevention and the Society of Cardiovascular Computed Tomography

SCCT/STR guidelines for coronary artery calcium scoring of noncontrast noncardiac chest CT scans: A report of the Society of Cardiovascular Computed Tomography and Society of Thoracic Radiology

Agreement of visual estimation of coronary artery calcium from low-dose CT attenuation correction scans in hybrid PET/CT and SPECT/CT with standard Agatston score

Association of Coronary Artery Calcification and Mortality in the National Lung Screening Trial: A Comparison of Three Scoring Methods

Ordinal scoring of coronary artery calcifications on low-dose CT scans of the chest is predictive of death from cardiovascular disease

Time Course of Lung Changes On Chest CT During Recovery From 2019 Novel Coronavirus (COVID-19) Pneumonia

Chest CT Features of COVID-19 in

The role of imaging in 2019 novel coronavirus pneumonia (COVID-19)

Antiplatelet therapy in valvular J o u r n a l P r e -p r o o f Journal Pre-proof and structural heart disease interventions

PCNA Guideline on the Management of Blood Cholesterol: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines

Atherogenic Lipoprotein Subfractions Determined by Ion Mobility and First Cardiovascular Events After Random Allocation to High-Intensity Statin or Placebo: The Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER) Trial

Statin use and high-dose statin use after ischemic stroke in the UK: a retrospective cohort study

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Estimation of cardiovascular risk on routine chest CT: Ordinal coronary artery calcium scoring as an accurate predictor of Agatston score ranges

Multiple Imputation for Nonresponse in Surveys

Boruta -A System for Feature Selection

Coronary Artery Calcification and Complications in Patients With COVID-19

Characteristics and outcomes of patients hospitalized for COVID-19 and cardiac disease in Northern Italy

Prevalence of Underlying Diseases in Hospitalized Patients with COVID-19: a Systematic Review and Meta-Analysis

Potential Effects of Coronaviruses on the Cardiovascular System

The Variety of Cardiovascular Presentations of COVID-19

Association of Treatment Dose Anticoagulation with In-Hospital Survival Among Hospitalized Patients with COVID-19

Pulmonary Vascular Endothelialitis, Thrombosis, and Angiogenesis in Covid-19

COVID-19 as a cardiovascular disease: the potential role of chronic endothelial dysfunction

Abdominal Imaging Findings in COVID-19: Preliminary Observations

Rationale for Consideration of Statins for COVID-19 Patients n

Statin Medication Enhances Progression of Coronary Artery Calcification: The Heinz Nixdorf Recall Study

Impact of statins on serial coronary calcification during atheroma progression and regression

Impact of Significant Coronary Artery Calcification Reported on Low-Dose Computed Tomography Lung Cancer Screening

Coronary Artery Calcification Is Often Not Reported in Pulmonary CT Angiography in Patients With Suspected Pulmonary Embolism: An Opportunity to Improve Diagnosis of Acute Coronary Syndrome

The authors thank Andrew Pagano MD, Sharon Steinberger MD, and