Key Points:
- Coronary artery calcium (CAC) scoring is widely used to screen for atherosclerotic coronary artery disease (ACAD). A CAC score of zero (CACS = 0) is often used to rule out significant coronary disease; however, non-calcified plaque may be present in 10–20% of individuals with CACS = 0, and whether this varies across diverse racial, ethnic, and geographic populations is uncertain.
- Using the GPS-CAD (Global Pretest Probability Study of Coronary Artery Disease) database of 32,690 participants from 15 countries who underwent both CAC scoring and coronary CT angiography (CCTA), investigators assessed the reliability of CACS = 0 for excluding coronary plaque across diverse populations.
- Among the 45% of participants with CACS = 0, 15% had plaque detected on CCTA (a “false negative” result). The combined variable of “Race/Ethnicity + Country” was the strongest predictor of both false negative and true positive CAC results, surpassing age, sex, and all traditional cardiac risk factors. False negative rates varied more than 4-fold by race/ethnicity (6% in Hispanic/Latino Whites to 21% in Middle Eastern individuals) and more than 8-fold by country (2% in Uganda to 28% in Egypt).
- These findings demonstrate that the ability of CACS = 0 to exclude atherosclerotic plaque varies substantially by race, ethnicity, country, and age, and that clinical decision-making regarding the use of CAC versus CCTA should take these factors into account.
Coronary artery calcium scoring (CACS) is a widely used, non-invasive screening tool for atherosclerotic coronary artery disease (ACAD). A CAC score of zero has traditionally been considered reassuring, and is used in clinical guidelines as a basis for deferring further testing in low-risk patients. However, prior studies have demonstrated that 10–20% of individuals with CACS = 0 may still harbor non-calcified coronary plaque, which can go undetected by CAC screening alone. This limitation may be particularly relevant in younger patients and certain racial or ethnic groups, but global data on the reliability of CACS = 0 across diverse populations have been lacking. Furthermore, existing pretest probability tools for coronary artery disease were largely developed in homogeneous, predominantly White Western populations, and their application to diverse global populations remains limited.
On March 28th, 2026, the results of “Reliability of CAC = 0 to Exclude Coronary Plaque in Diverse Global Cohorts: A Global Probability Study of Coronary Artery Disease (GPS-CAD) Analysis” were presented at ACC Scientific Sessions 2026, with simultaneous publication of the GPS-CAD study design and rationale in JACC: Asia. The purpose of this analysis was to assess the presence and predictors of atherosclerotic plaque in individuals with CACS = 0 across racially, ethnically, and geographically diverse groups.
GPS-CAD is a multicenter retrospective cross-sectional observational study that aims to collect real-world data on the global patterns and predictors of early-phase ACAD [NCT05722145]. The current analysis included 32,690 participants from 15 countries across 22 populations (encompassing 6 continents) who underwent both CAC scoring and CCTA. Data were collected on demographics, cardiovascular risk factors, race, ethnicity, and medications. Three primary ACAD endpoints were assessed: CACS > 0, any coronary plaque on CCTA (>0% stenosis), and obstructive CAD (≥50% stenosis). For this analysis, scan findings were classified as true negative (CAC–/CCTA–), false negative (CAC–/CCTA+), true positive (CAC+/CCTA+), or false positive (CAC+/CCTA–). Covariate relative importance was assessed by AUC, and multivariable models were adjusted for age, sex, body mass index, symptom status, diabetes, hypertension, dyslipidemia, smoking, family history, and originating study design (e.g. registry, trial, or clinical).
The mean age of participants was 59 years, 45% were female, and 73% were symptomatic. In terms of race & ethnicity, 39% were Non-Hispanic/Latino White, 29% were Hispanic/Latino White, 20% were Asian, 5% were Black, and 4% were Middle Eastern. Overall, 55% had a CACS > 0, 60% had any plaque on CCTA, and 18% had obstructive CAD.
CACS = 0 was present in 14,807 participants (45%), of whom 12,525 (85%) were true negatives and 2,282 (15%) were false negatives (e.g., plaque present on CCTA despite CACS = 0). The combined variable of “Race/Ethnicity + Country” was the single strongest predictor of both false negative results (AUC 0.64) and true positive results (AUC 0.84), outperforming age, sex, and all traditional cardiac risk factors.
The prevalence of false negative CAC results varied substantially by country, ranging from 2% in Uganda to 28% in Egypt. In multivariable models, country was an independent predictor of false negative results: compared to the USA (NPV 0.98), the adjusted odds of a false negative ranged from 1.3-fold to 8.9-fold higher across other countries, with Egypt having the highest false negative rate (aOR 8.86 [95% CI: 6.91, 11.37]; NPV 0.53). Similarly, false negative rates varied by race/ethnicity, from 6% in Hispanic/Latino Whites to 21% in Middle Eastern individuals. In multivariable models, race/ethnicity was also an independent predictor: compared to Non-Hispanic/Latino Whites, the adjusted odds of a false negative ranged from 1.3-fold (Hispanic/Latino Whites; aOR 1.32 [95% CI: 1.13, 1.54]) to 5.0-fold (Middle Eastern; aOR 5.04 [95% CI: 4.17, 6.10]). Younger age (<55 years) was also independently associated with false negative results (aOR 1.36 [95% CI: 1.24, 1.50]). Notably, no independent differences in false negative rates were observed based on sex, symptom status, or diabetes.
Limitations of this study include its retrospective, cross-sectional design using pooled data from clinical registries, randomized trials, and observational studies, each with their own inclusion criteria and potential selection biases. Differences in access to cardiac CT across countries may have affected the study population. The classification of race/ethnicity was not standardized across sites, as there is no universal classification system for these variables in clinical research. Additionally, CT interpretation was based on site reports rather than a central core laboratory for most cohorts, though core laboratory reports were prioritized when available.
Lohendran Baskaran, MBBS, of the National Heart Centre Singapore and Duke-NUS Medical School, concluded: “In patients with suspected coronary artery disease, decision-making regarding the use of CAC or CCTA, and interpretation of CACS = 0 should consider race, ethnicity, country, and age. The use of CACS = 0 to exclude plaque should be applied with caution in diverse groups.”