Low or Even Zero CAC may not be Warranted as Safe in Thai Officers: A 10-year prospective cohort study

OBJECTIVES: CAC has been widely accepted as a strong independent event predictor in asymptomatic moderate risk in the western population. However, the role of coronary artery calcium (CAC) has not been established in the Thai population yet. Thus, we conducted a 10-year prospective study to verify the predictive role of CAC in asymptomatic Thai participants. MATERIAL AND METHODS: A total of 239 asymptomatic officers were voluntarily enrolled and consented to take part. We excluded half (51.5%) of them had no detectable calcified coronary plaque using the Agaston-130 method (CAC = 0). There was no statistical difference in age, gender, SBP, fasting blood sugar (FBS), total serum cholesterol (T. Chol), serum TG, HDL-C and cigarette smoking between candidates with absent or present CAC, except the mean LDL-C which was statistically significantly higher in the group with no CAC, 154.7 vs 143.7 mg/dl, p = 0.036. After ten years, there were 23 CVD events including 3 ACS cases, 9 strokes (all ischemic type) and 11 deaths (91% were non-CVD causes). There was no statistically significant difference between these events in the two groups, CAC = 0 (n = 121) vs CAC > 0 (n = 108) regarding to; ACS 0.8% vs 1.9% ( p = 0.603), stroke 5% vs 2.8% ( p = 0.506), death 2.5% vs 7.4% ( p = 0.121) and the combined outcome (ACS + stroke + death): 8.3% vs 12% ( p = 0.384). The mean CAC in CVD victims was quite low, 32.76 (0 - 53.1) for three ACS cases and 23.75 (0 - 180.9) for nine strokes. CONCLUSION: While the small number and the prominent male gender of candidates were the major limitations and precluded extrapolation to others, this study was the first 10-year cohort that reported different results from western studies, low or even zero CAC might not be safe in asymptomatic Thai people. Thus, we should be skeptical to use CAC as a CVD risk predictor in our population. Whether or not Thai men had low prevalence of calcified coronary plaque or had a different pathophysiologic mechanisms, further study in a larger population is mandatory.


Figure 1:
The study flow chart is illustrated. After exclusion, a total of 229 cases underwent CAC scan. C ardiovascular diseases (CVD) remained the leading causes of death worldwide and accounted for one-third of the global death. 1 In 2016, 17.9 million people died from CVDs and 85% of deaths were from heart attack and stroke. 2 In China, stroke and coronary artery disease (CAD) are the first and second leading cause of CVD death. 3,4 In Thailand, reports from the Ministry of Public Health indicated that more Thai people had died from CVD than in the previous decade. From 2007-2014, the rate of sickness from CAD and stroke rose 24% and 41% respectively. 5 The mortality rate of CVD increased from 21.3 to 28.9:100,000 (2008-2014) for CAD and from 20.8 to 31.7:100,000 (2008-2012) for stroke. 5,6 It is clear that the population at risk needs to be identified to prevent future CVD events.
CVD risk assessment by traditional risk factors has been generally recommended to predict the 10-year CVD risk by current practice guidelines. 7,8 However, the predictive accuracy for the future CVD events by traditional risk score i.e. Framingham score, was only modest, in the range of 50-65%. 9 Coronary artery calcium (CAC) is the consequence of atherosclerotic healing process so it can be used to quantify the amount of calcified plaque for risk stratification. Several research studies have verified the role of CAC in predicting hard CVD events in asymptomatic patients and following the progression of the atherosclerotic process. CAC is proven to be a better risk estimator, beyond systemic risk scores, in predicting coronary events, stroke, and heart failure. [10][11][12][13][14][15][16] However, all of these evidences come from western population studies and the role of CAC remains undocumented in Thai people. Thus, we conducted a 10-year prospective study in asymptomatic Thai officers who underwent CAC screening in 2006 and followed their CVD events through 2016.

Material and Methods
After receiving approval by the hospital ethical committee, this study was initiated in 2006 during the annual physical check-up. Only asymptomatic candidates aged between 35-60 years were voluntarily enrolled. (see Figure 1). We excluded officers who had prior CVD, were not willing to sign the informed consent for participation and pregnant candidates. The family history of premature atherosclerosis, dyslipidemia, cigarette smoking, blood pressure, lipid profiles and blood glucose were obtained. 17,18 To rule out pre-existing CAD, all participants underwent exercise stress test by the standard Bruce protocol. Candidates who had ischemic response within 7 METS would be referred for coronary angiography (CAG). Coronary revascularization or medication were justified based upon the result of individual angiographic results. Candidates with documented CAD were later excluded for outcome measurements. The study flow is illustrated in Figure 1.
CAC study: CAC volume is obtained by using 256 -slice MDCT, Philips system and measured by Agaston method which defines 130 HU or more for threshold to identify CAC. CT scan for CAC is operated by non-contrast technique at tube current 55 -80 mAs with tube voltage of 120 kVp. Based upon the CAC volume, the population studied is divided into two groups, CAC = 0 versus CAC > 0. The occurrence of CVD events in all candidates were followed up annually for consecutive ten years (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016) Clinical end points: Candidates who had positive EST were excluded. The primary end points were fatal and non-fatal ACS and any types of stroke. Secondary end point was all causes of mortality.

Statistical analysis:
The Student t test and the Pearson's Chi-square were used to compare the different outcomes between candidates with and without CAC, and p < 0.05 was considered statistically significantly different.
By CT CAC with Agaston-130 method, half of candidates (51.5%) had no detectable calcified plaque, CAC = 0. The Abbreviation: SBP = Systolic blood pressure, FBS = Fasting blood sugar, T.chol = total serum cholesterol, TG = serum triglyceride, HDL-C = High density lipoporotein cholesterol, LDL-C = Low density lipoprotein cholesterol, *p < 0.05 Low or Even Zero CAC may not be Warranted as Safe in Thai Officers: A 10-year prospective cohort study demographic of candidates between zero CAC score (CAC = 0) and more than zero (CAC > 0) was shown in Table 1. There was no statically significant difference in gender, age, mean values SBP), FBS, T.Chol, TG), HDL-C and present-day cigarette smoking. However, the mean LDL-C of candidates with zero CAC was significantly higher than that of the CAC > 0 candidates, 154.7 vs 143.7 mg/dl, p = 0.036.
Exercise stress test (EST) result: Ten cases (4.2%) had ischemic EST response within 7 METS and underwent coronary angiography. All of them had explainable causes of exercise induced myocardial ischemia. Two cases from CAC absent group had myocardial bridging and coronary slow flow (CSF). Eight cases were from the CAC present group (CAC > 0), five cases (62.5%) had severe CAD and one of each other three cases had coronary spasm, myocardial bridging and CSF. In five cases with severe CAD, three cases had percutaneous coronary stent implantation, one case had bypass surgery and another one needed aggressive medication. (Figure 3). The calcified plaque distribution and coronary angiographic findings are illustrated in Table 2. Since severe CAD, coronary spasm, CSF and myocardial bridging could result in either ACS or sudden death, all of these ten positive EST tests were excluded for outcome calculation.  Table 1: Demographic data and conventional risk factors of candidates with detectable calcified plaque (CAC > 0) and absent calcified plaque (CAC=0) are displayed in proportion (%) and mean ± standard deviation in mg/dl. There was no statistically significant difference except the mean LDL-C which was significantly higher in candidates with absent CAC.    Clinical end points: after ten years, there were a total of 23 events including 3 ACS cases, 9 ischemic strokes and 11 deaths, see Table 3. The three ACS cases appeared at the mean duration of six years and their average age was 57.7 years. Their angiographic findings are illustrated in Figure 4. Stroke cases appeared later than ACS, within 7.5 years, and they were older (mean age was 61.8 years). Eleven cases (mean age was 55.3 years) had died with the average time of 6.45 years. The major causes of death were non-CVD death (91%), including cirrhosis 36.4%, cancer 18.2% and the rest, 9.1% (one of each expired from immune deficiency, sepsis, drowning and accident). Only one CVD death case occurred at age 63 years. He had acute basal ganglia infarction and later expired from heart failure within a few weeks.
The amount of calcified plaque (CAC), measured by the Agaston method of ≥ 130 HU, is displayed in range, from 0, 1-100, 101-200, 201-300, 301-400 and > 400, and correlated with the clinical outcomes in Table 4. It was noted that calcified plaque burden, CAC < 100, was detected in 100% of ACS cases, 89% of stroke victims and 54.5% of death patients. There was no statistical difference between events and CAC severity.  Table 3 : Clinical outcome between candidates without (CAC = 0) and with calcified coronary plaque (CAC > 0). There was no statistically significant difference between the two groups.

CAC Zero is safe in western population
Calcified coronary artery plaque burden, so-called CAC, is a well-established independent marker for predicting cardiovascular events and death in western population. [11][12][13][14][15][16][19][20][21] The absence of CAC, zero CAC, has been associated with a very low risk of having CVD events 13,14,22 and all causes mortality. 23 (Table 5). In the longest prospective study of 14.6 years, involving 9,715 individuals, the overall mortality of 4,864 individuals with CAC = 0 group was three times lower than that of CAC > 0 group, (4.7% vs 14.6%). 24 Thus, the author concluded that CAC was the strongest mortality predictor with the hazard ratio of 2.67 (95% CI 2.39-3.11, p < 0.05). 24 Currently, CAC had been recommended by ACC/AHA preventive guideline for individual risk assessment 11,25 and guiding statin therapy. 26 From the American Heart Association (AHA)/American College of Physician (ACC) recommendation in 2018, adults 40 to 75 years of age without diabetes mellitus and with LDL-C levels ≥ 70 to 189 mg/dL, at a 10 -year ASCVD risk score of ≥ 7.5% to 19.9%, treatment with statin therapy may be withheld or delayed if CAC is zero, except in cigarette smokers, those with diabetes mellitus, and those with a strong family history of premature ASCVD. 26

Exclusion of occult CAD in asymptomatic candidates
Since the predictive role of CAC had not been documented in the Thai population, CAC screen was performed in 239 volunteered officers who had no history of CVD. To further exclude occult CAD, EST was performed in all candidates and ten cases (4.2%) had ischemic exercise response. Half of them, five cases, had occult severe coronary stenosis that required either revascularization (four cases) or aggressive medication in another (Table 2 and Figure 3), so these cases were excluded. The other half (five cases) had potential causes of myocardial ischemia including coronary slow flow in three cases, vasospasm in one case and myocardial bridging in another. Since life-threatening arrhythmias, myocardial infarction or even sudden death had been previously reported with CSF 27-29 , vasospasm 30-32 and myocardial bridging 33-38 cases, we also excluded them from statistical analysis in order to prevent confounding clinical outcomes.

CAC Zero may not be safe in Thais
Despite providing healthy life style education and medication, 21 events still appeared overall within ten years. There was no statistically significant difference between candidates of CAC 0 vs CAC > 0, regarding ACS: 0.8% vs 1.9% (p = 0.603), stroke: 5% vs 2.8% (p =0.506), total death: 2.5% vs 7.4 % (p = 0.121) and the combined outcome (ACS + stroke + death): 8.3% vs 12%, p = 0.384, see Table 3. Our study did show different results from western data that CAC zero might not have definite power to protect ACS, stroke and overall death in our population. Thus, CAC zero, only by itself, could not yet be proven as a sole protective indicator for low CVD risk in asymptomatic Thai patients. From Table 4, there is no statistically significant difference in CAC severity and clinical outcome including EST results. In addition, the mean CAC in all vascular cases was quite low; 32.76 (0 -53.1) for three ACS cases and 23.75 (0 -180.9) for nine stroke patients. By AHA/ ACC guidelines 11,25 , all ACS cases and 88.9% of stroke victims would be classified as low risk candidates since their calcified plaque burden was low, below 100.

How to explain the low CAC and events
Racial difference in CAC has been addressed in previous studies. [36][37][38][39][40] The first autopsied study that showed racial difference of calcification in major coronary arteries was reported in 1965. 39 Of a total of 777 autopsied cases that died from non-coronary causes, the prevalence of calcified coronary arteries was three times less in black patients, in comparison to white cases, 20% vs 75% . 40 In a MESA study, the highest Chaothawee L, et al. CAC = Coronary artery calcium, CVC = Central venous catheter, CAD = Cardiovascular disease, CHD = Coronary heart diseases, ACS = Acute coronary syndrome CAC was observed among asymptomatic white subjects and the lowest value was also found in the black candidates. 40 In the elderly, Chinese American men (age >70 years) and women (aged >75 years) had the lowest CAC. 41 45 These studies 44,45 supported our findings that the Asian population tends to have lower coronary calcified plaque burden than Caucasian populations. It also suggested that, without a population-based study or adjusting the CAC cut point, low risk candidates (CAC = 0), by the western guideline, may not be simply applicable to the Thai population. There are several possible causes of low CAC in Asian populations including genetic, behavioral or cultural factors (food, exercise etc.), and all of these remain to be verified.
It is worth mentioning that zero CAC does not mean the absence of calcified plaque since the cut-off point is an arbitrary limit set by the Agaston-130 method. 11 A negative test (score 0) does not exclude the lipid rich plaque that might be prone to rupture in the presence of a low and undetected calcified healing process. In the final MESA study, involving 3,923 asymptomatic cases, no detected CAC was found in 3,415 cases (87%) and there were 28 CVD events (0.82%) which was three-fold lower than those in the CAC positive group. 46 The associated risk factors in the negative CAC were prior smoking (HR = 3.57; 1.08-11.77), current smoking (HR = 4.93; 1.20-20.30), and diabetes (HR=3.09; 1.07 -8.93) (43). None of our three ACS cases were diabetic but all of them were current smokers and had a high LDL-C, mean LDL-C of 182.67 mg/dl. One case with zero CAC not only smoked cigarettes, but also had premature atherosclerosis in a family member. These factors have been known to be associated with an athero-thrombotic process [47][48][49] and all negative CAC cases should be looked for these condition before justifying a future CVD risk.

Conclusion
Several limitations existed in our study, first, the small number of cases from a single center which might have less statistical power to justify the predictive role of CAC. Second, the population study was Thai officers, aged 35-60 years, and the majority of them were men, which precluded application to other population groups. Third, all of them were medically treated and educated for lifestyle modification. However, all candidates were voluntarily recruited without compensation, so this study should have less selection bias in this regard. Based on this result, patients with low or even zero CAC could not be viewed as low risk cases and zero CAC did not warrant as a safety marker at least in Thai officers. Further study in a larger population is mandatory to accurately determine the predictive role of CAC in our population.
As mentioned in discussion section, the cut-off point of CAC is set by the Agaston-130 method. To reduce CAC detection threshold to detect microcalcification (which is undetected with cutoff point Agaston 130) may be possible. However, reducing CAC detection threshold would make the CAC scoring susceptible to image noise which may affect to the accuracy in CAC scoring. In addition, no available supporting data of improvement risk prediction yet by improved microcalcification detection method. 50