Risk stratification using coronary artery calcium and potential benefit of semaglutide therapy: A cost‐effectiveness modelling study

Abstract

Aim

Coronary artery calcium (CAC) scoring is observed to improve risk stratification for major adverse cardiovascular events (MACE). Semaglutide, a recently introduced anti‐obesity drug, is very effective, but wider use is limited due to high costs. Hence, this study investigated the cost‐effectiveness (CEA) of semaglutide across CAC groups.

Materials and Methods

CAC scores for 38 058 CLARIFY registry participants meeting SELECT criteria were included. They were stratified into four CAC groups: 0, 1–99, 100–399, ≥400. To determine MACE (composite of myocardial infarction, heart failure, stroke or all‐cause mortality) risk across CAC groups, hazard ratios (HR) were estimated from multi‐variable adjusted Cox proportional hazard models. Next, lifetime‐horizon Markov models were created to simulate semaglutide therapy and the potential clinical benefit (reported as number needed to treat [NNT]) and CEA (estimated with incremental cost‐effectiveness ratio [ICER]) were examined for CAC groups. Multiple scenarios to mimic real‐world experience were fitted for robust sensitivity analyses.

Results

Compared to CAC = 0, MACE risk was higher for CAC ≥400 (HR: 1.97 [95% CI: 1.66–2.35]), heart failure (HR: 1.76 [95% CI: 1.36–2.28]), mortality (HR: 1.62 [95% CI: 1.21–2.17]). Modelling 3.3 years of semaglutide use resulted in potential MACE NNT values of 151 (95% CI: 108–302) and 34 (95% CI: 25–69) for CAC = 0 and CAC ≥400. Markov modelled ICER for semaglutide use reduced across CAC groups ($625 863/QALY [CAC = 0] vs. $168 666/QALY [CAC ≥400]).

Conclusions

CAC scores have potential use as a tool to estimate potential clinical benefit and cost for lifetime semaglutide therapy among obese individuals.

1. INTRODUCTION

The prevalence of obesity continues to rise worldwide; a recent study ¹ projected that by 2030 close to half the US adult population would be obese while another recent study reported that, by 2050, 213 million adults in the US would be obese. ² , ³ Obesity has many downstream effects on health, particularly cardiometabolic health and may even reduce the individuals' lifespan. ⁴ The recently reported SELECT trial supported the cardiovascular benefit of semaglutide 2.4 mg in obese individuals; over a 3.3 year median follow‐up time, the trial reported a 20% relative risk reduction for major adverse cardiovascular events (MACE) (HR: 0.80, 95% CI: 0.72–0.90) for MACE. ⁵ Positive benefits were also observed for incident heart failure (HR: 0.82, 95% CI: 0.71–0.96) and all‐cause mortality (HR: 0.81, 95% CI: 0.71–0.93). However, despite these proven benefits, wider use of semaglutide remains limited due to high costs, insurance denials, and supply constraints. ⁶ , ⁷ Prior modelling studies evaluating semaglutide have primarily focused on trial‐based populations or aggregate cardiovascular risk estimates and have not integrated imaging‐based risk stratification to inform lifetime cost‐effectiveness. In particular, to our knowledge, no prior analyses have evaluated coronary artery calcium (CAC) as a modifier of projected clinical benefit and economic value of semaglutide therapy. The present study extends existing work by combining real‐world event rates from a large observational registry with a Markov‐based economic model to examine how CAC burden may identify subgroups in whom semaglutide therapy yields relatively greater clinical and economic value.

Traditionally, models such as the Predicting Risk of Cardiovascular Disease EVENTs (PREVENT) score are used to predict incident MACE. ⁸ Although the PREVENT equations demonstrate only moderate discrimination, with C‐statistics in the range of approximately 0.65, ⁹ prior evaluations have shown good calibration and overall performance comparable to or exceeding traditional pooled cohort equations. Moderate discrimination is common among cardiovascular risk prediction models, particularly in heterogeneous populations. Accordingly, CAC may offer complementary value by refining risk stratification beyond a calibrated clinical risk score, particularly in identifying individuals at the extremes of long‐term cardiovascular risk. Additionally, a recent editorial on the PREVENT calculator suggested that adding the CAC score may improve the discrimination and calibration metrics of this risk prediction model. ¹⁰ CAC, quantified via non‐contrast cardiac computed tomography (CT), is increasingly becoming a favoured tool to determine incident MACE risk. Termed ‘the power of zero’, CAC has been reported to have a high and reliable negative predictive value. ¹¹ Conversely, a raised CAC score may denote the possibility of sub‐clinical coronary artery disease and these individuals may benefit from aggressive preventive therapies such as semaglutide. ¹² In fact, analysis of MESA data also reported that a 5‐unit annual increase in CAC scores in those with a baseline CAC = 0 led to a 40% relative increase in coronary heart disease. ¹³

Hence, this study used real world data from registry participants that were eligible to receive semaglutide therapy for treating obesity and investigated the following: (i) Were higher CAC scores at registry enrollment associated with more MACE events? (ii) Is there a benefit of risk‐stratification based on CAC scores and using this to decide semaglutide therapy? and lastly (iii) What would be the cost‐effectiveness of such a strategy?

2. METHODS

This study was approved by the ethical board of the University Hospitals Healthcare System as part of the wider CLARIFY registry. Individuals also provided consent for research at the time of their enrollment into the CLARIFY registry. The study was conducted and reported as per the STROBE guidelines and economic evaluation was conducted and reported in accordance with the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) guidelines (checklist provided in the Supporting Information).

2.1. CLARIFY registry

The CLARIFY registry (ClinicalTrials.gov NCT04075162) was established within University Hospitals Health System (UHHS), one of the largest integrated healthcare networks in northeast Ohio, comprising 11 hospitals and over 31 health centres. ¹⁴ In response to the lack of Medicare reimbursement for CAC testing in Ohio, UHHS launched a system‐wide low‐cost CAC screening programme in 2014, offering scans at $99. A temporary no‐charge pilot was implemented in June 2015 to assess community uptake, followed by full implementation of a no‐charge CAC programme in January 2017. CAC screening was offered to men aged ≥45 and women aged ≥55 without known cardiovascular disease but with at least one traditional risk factor (e.g., dyslipidaemia, hypertension, smoking, diabetes, or a family history of premature coronary artery disease), and to men and women aged ≥40 with chronic inflammatory conditions (e.g., inflammatory bowel disease, lupus, rheumatoid arthritis, ankylosing spondylitis, psoriasis). Testing was conducted at 21 radiology sites across northeast Ohio. For the present analysis, we included all individuals who underwent CAC screening within UHHS between 1 January 2014 and 31 December 2023. As shown in Figure S1, we determined eligibility for semaglutide use as treatment for obesity according to the inclusion criteria of the SELECT trial to 111 043 participants of the CLARIFY registry to develop a cohort of 38 058 participants that are the cohort of interest in this study. Unlike the study cohort in the SELECT trial, the CLARIFY registry does not contain patients with established cardiovascular disease. However, despite this, we carefully selected people from this registry who would be eligible to receive semaglutide therapy as treatment for obesity according to trial criteria. We identified participants in the CLARIFY registry with body mass index ≥27 kg/m² and without diabetes, consistent with key phenotypic features of the SELECT trial. However, CLARIFY is a primary prevention registry and does not include individuals with established cardiovascular disease at enrollment. Accordingly, this analysis should be interpreted as a modelling study evaluating the potential role of CAC for risk stratification and projected cost‐effectiveness of semaglutide therapy in a primary prevention population, rather than as a direct replication or extension of the SELECT trial.

2.2. Endpoints

The median (maximum) follow‐up period for the study was 2.3 (9.9) years. The primary endpoint for our study was MACE defined as a composite of nonfatal myocardial infarction, non‐fatal stroke and all‐cause mortality. This modification was done as unlike the SELECT trial, our registry did not contain cause of death and hence we were unable to identify cardiovascular mortality. However, as described below, the relative benefit for this revised MACE endpoint (including all‐cause rather than cardiovascular mortality) was derived using trial information. Secondary endpoints included in our study were incident heart failure (HF) and all‐cause mortality. Non‐fatal events were identified using the International Classification of Disease (ICD) 10th edition codes (MI—I21x, I22x, I23x, I24.1, I25.2; stroke—I63x, I64x; heart failure—I50x).

3. STATISTICAL ANALYSIS

Our cohort of 38 058 adults was grouped based on their enrollment CAC scores into the following groups: CAC = 0, CAC 1–99, CAC 100–399, and CAC ≥400. Missing values in covariates were addressed using multiple imputation by chained equations. We generated m = 20 imputed datasets using regression models appropriate to variable type and combined estimates across imputations using Rubin's rules. The extent of missingness for key variables is summarized in the Supporting Information. Baseline characteristics across CAC groups were compared with the Kruskal‐Wallis rank sum test or χ² test as appropriate. Multivariable Cox proportional hazards models were constructed to evaluate the association between CAC score and both primary and secondary endpoints. These models were adjusted for the following demographic and clinical covariates: age at enrollment (years), sex, race/ethnicity, body mass index (BMI—kg/m²), systolic blood pressure (mmHg), smoking, HbA1c level (%), low density lipoprotein cholesterol level (mg/dL), current statin therapy, and current blood pressure medication use. As a sensitivity analysis, matched cohorts of CAC groups were created using propensity scores. Further details regarding the procedures used or matching are provided in the Supporting Information (Matched Cohort). Stratified cox proportional hazard models were then fitted to this data to investigate the risk for MACE in CAC groups 1–99 100–399, and ≥400 with CAC = 0 as reference.

To model the potential benefit of semaglutide, hazard ratios (HR) observed in the SELECT trial were obtained. To account for the fact that our MACE endpoint included all‐cause mortality rather than cardiovascular mortality (which was included in the SELECT trial), HR for MI, stroke, and all‐cause mortality from the SELECT trial were pooled using a random‐effects meta‐analysis with restricted mean likelihood to derive the HR for this composite endpoint of MACE that included all‐cause mortality rather than cardiovascular mortality (Figure S2). To model variation in potential clinical benefit (stochastic uncertainty), a Monte Carlo simulation approach was used, and individual HR were drawn at random from a triangular distribution. This statistical modelling was applied to all participants in each CAC group over a hypothetical 3.3 year (median follow‐up time of the SELECT trial) and 5‐year time frame. Using the observed events as the baseline, the absolute risk reduction of semaglutide use for each CAC group was estimated and translated into number needed to treat (NNT) at 3.3 years (NNT_3.3) and 5 years (NNT₅). A dose–response curve was also constructed to illustrate the relationship between the CAC score as a continuous variable and the hypothetically obtained NNT_3.3 and NNT₅.

3.1. Cost effectiveness analysis

Markov models on an annual time‐scale were constructed to calculate the CEA of semaglutide versus usual care from a US healthcare payer perspective over a 40‐year time horizon. A deterministic model was used to estimate the incremental cost‐effectiveness ratio (ICER) for semaglutide use for each CAC group incorporating quality‐adjusted life years (QALYs), costs, and event probabilities into the Markov models. Health states included no prior cardiovascular event, post–myocardial infarction, post‐stroke, and death (absorbing). Transition probabilities between health states were estimated separately by CAC stratum using observed event rates from the CLARIFY registry. Cost inputs included drug costs, cardiovascular event costs, and follow‐up care, while QALY estimates incorporated modelled survival and utility decrements for MACE (MI, stroke, all‐cause death) events (Table S1). Costs were modelled using gamma distributions, health state utilities using beta distributions, and relative risks using log‐normal distributions, consistent with established health economic guidelines. Both costs and health outcomes were discounted at an annual rate of 3%. A half‐cycle correction was applied to account for transitions occurring, on average, midway through each cycle. Model assumptions and face validity were assessed through internal consistency checks and review by clinical and health economic experts. These parameters were used by us and others in prior cost‐effectiveness models. ¹⁵ , ¹⁶ The annual inflation adjusted semaglutide cost for our primary Markov model was fixed at $13.937.88 which is the mean of the maximum and minimum value of published US costs ($8600–$17 597). ¹⁷ , ¹⁸ Probabilistic sensitivity analysis (PSA) with 1000 random draws was fitted to account for uncertainty in costs, utilities, and event probabilities. Parameter distributions for this model were based on published literature and our cohort information. PSA Results were summarized using mean, median incremental costs and mean, median incremental QALYs, and decision uncertainty was evaluated using cost‐effectiveness acceptability curves across a range of willingness‐to‐pay thresholds. To model real‐world situations (I) Drug discontinuation: the impact of drug discontinuation was studied where 30% semaglutide users stopping taking the medication after the first year and 5% in each subsequent year over the first 10‐years of the Markov model, (II) Future Semaglutide cost: Considering the possibility that semaglutide costs in the US will reduce over the 40‐year Markov model, analyses were conducted incrementally reducing semaglutide annual cost by $500 from the current cost to a minimum cost of $2000.

4. RESULTS

4.1. Study cohort

From the CLARIFY registry, 38 058 (mean age was 58 years, 49% females) were included in this study. At enrollment, a CAC score = 0 was observed in 16 624 (44%), CAC score 1–99 in 11 262 (30%), CAC score 100–399 in 5843 (15%), and a CAC score ≥400 was observed in 4329 (11%) participants. Older age, male sex, lower HDL, higher systolic blood pressure, reduced kidney function, statin use, and antihypertensive therapy were likely to have higher CAC scores (Table S2). Baseline characteristics for the matched cohort are presented in Table S3. Overall, 1707 MACE events (4.5%), 819 HF events (2.2%), and 642 deaths (1.7%) occurred over the study period. Observed MACE, HF and mortality events and rates in our cohort incrementally increased across CAC groups (Figure 1 top panel, Table 1). Adverse cardiovascular event rates (MACE, HF, mortality) were also found to increase incrementally across CAC groups in the matched CAC groups (Figure 1 bottom panel). In adjusted multi‐variable models, with CAC = 0 as the reference, the relative risk for MACE (HR: 1.97, 95% CI: 1.66–2.35), HF (HR: 1.76, 95% CI: 1.36–2.28), and all‐cause mortality (HR: 1.62, 95% CI: 1.21–2.17) were higher in those with CAC ≥ 400 (Table S4).

FIGURE 1.

Adverse event rates and hypothetical number needed to treat for semaglutide therapy by CAC category. The left panel shows the incidence rates (per 1000 patient‐years) for MACE, HF, and all‐cause mortality in the CLARIFY cohort stratified by CAC category. The right panel depicts the hypothetical benefit of Semaglutide therapy expressed as the number needed to treat (NNT) to prevent one event over the study horizon. Maroon gradient bars represent estimates from the CLARIFY cohort, and grey bars represent the corresponding estimates from the SELECT trial. For both panels, numerical values with 95% confidence intervals are displayed above each bar. CAC, coronary artery calcium; HF, heart failure; MACE, major adverse cardiovascular events.

TABLE 1.

Adverse events observed during the study period.

Whole cohort
Endpoint	Overall (N = 38 058)	CAC = 0 (N = 16 624)	CAC 1–99 (N = 11 262)	CAC 100–399 (N = 5843)	CAC ≥400 (N = 4329)
MACE, n (%)	1707 (4.5)	425 (2.6)	470 (4.2)	388 (6.6)	424 (9.8)
Heart failure, n (%)	819 (2.2)	209 (1.3)	215 (1.9)	180 (3.1)	215 (5.0)
Mortality, n (%)	642 (1.7)	162 (1.0)	162 (1.4)	145 (2.5)	173 (4.0)

Matched cohort
	Overall (N = 5097)	N = 1292	N = 1263	N = 1278	N = 1264
MACE, n (%)	300 (5.9)	54 (4.2)	62 (4.9)	82 (6.4)	102 (8.0)
Heart failure, n (%)	150 (2.9)	28 (2.2)	31 (2.4)	39 (3.0)	52 (4.1)
Mortality, n (%)	112 (2.2)	22 (1.7)	22 (1.7)	29 (2.3)	39 (3.1)

Note: This table presents the number of adverse cardiovascular events observed in our study cohort during 17.9 per 1000 patient‐years of follow‐up. As reported, the number of adverse events increased incrementally across CAC groups. Figure 1 presents the corresponding event rate per 1000, person‐years of follow‐up.

4.2. Hypothetical clinical benefit

Applying SELECT trial risk reductions to observed event rates, we found that absolute risk reduction from modelled semaglutide therapy increased incrementally across CAC groups. The modelled NNT_3.3 for MACE (a composite of MI, stroke, all‐cause mortality) was 151 (95% CI: 108–302) in CAC = 0 versus 34 (95% CI: 25–69) in CAC ≥400 (Figure 2, Table S4). Similarly, the modelled NNT_3.3 for HF and mortality were 411 (255, 1852) and 359 (235, 974), respectively, for CAC = 0 versus 84 (52, 379) and 80 (52, 217) for CAC ≥400 (Figure 2 top panel). When similar analyses were done on the matched cohort, this same incremental benefit for semaglutide therapy was observed across CAC groups for all three adverse events investigated in our study (Figure 2 bottom panel). Similar incremental benefits were observed when the modelled NNT were calculated at the 5‐year follow‐up time point (Table S5). Models fitted to both the whole and matched cohorts using the CAC score as a continuous variable also demonstrated the incremental reduction in the NNT (both NNT_3.3 and NNT₅) with rising CAC values (Figure 3).

FIGURE 2.

Hypothetical number needed to treat (NNT) modelled for semaglutide therapy. These fan plots present the hypothetical NNT obtained when CAC score was considered as a continuous variable and Semaglutide therapy was modelled over a 3.3‐year (NNT_3.3) and 5‐year (NNT₅) time period. CAC, coronary artery calcium; HF, heart failure; MACE, major adverse cardiovascular events.

FIGURE 3.

Incremental cost‐effectiveness ratio (ICER) according to varying annual semaglutide cost and cost‐effectiveness of semaglutide therapy across CAC categories. The left panel shows the incremental cost‐effectiveness ratio (ICER) of semaglutide therapy as a function of annual drug price over a 40‐year time horizon. The dashed horizontal line denotes the willingness‐to‐pay (WTP) threshold of $50 000 per QALY. At this threshold, the annual semaglutide price yielding cost‐effectiveness is approximately $4000 (€3680). The right panel displays cost‐effectiveness acceptability curves (CEACs) illustrating the probability that Semaglutide is cost‐effective compared with usual care across a wide range of WTP thresholds. CEACs are stratified by coronary artery calcium (CAC) category, and frontier points represent the most cost‐effective strategy at each WTP value. Probabilities reflect results from probabilistic sensitivity analysis (PSA; 2000 simulations). CAC, coronary artery calcium; CEAC, cost‐effectiveness acceptability curve; ICER, incremental cost‐effectiveness ratio; PSA, probabilistic sensitivity analysis; QALY, quality‐adjusted life year, conversion ratio 1 USD = € 0.92 (obtained on October 27th, 2025); QALY, quality‐adjusted life‐year; WTP, willingness‐to‐pay.

4.3. Cost‐effectiveness

Deterministic modelling demonstrated that the ICER for semaglutide versus usual care decreased substantially across CAC groups (Table 2). For CAC = 0, the ICER was $625 863/QALY, compared with $425 492/QALY for CAC 1–99, $296 505/QALY for CAC 100–399, and $168 666/QALY for CAC ≥400. Probabilistic sensitivity models reported a mean ICER of $633 833/QALY for CAC = 0, $431 258/QALY for CAC 1–99, $301 186/QALY for CAC 100–399, and $171 113/QALY for CAC ≥400. Sensitivity analysis conducted to model drug discontinuation also reported benefit for semaglutide use albeit with the same high ICER observed in the primary analysis (Table S6). As reported in Figure 3, in the highest risk group (CAC > 400) the annual cost for semaglutide needs to reduce to $4000 (~ €3680) in the US for the ICER to reach $50 000/QALY.

TABLE 2.

Results of the cost‐effectiveness model.

CAC group	Usual care (QALY)	Usual care + semaglutide (QALY)	Usual care (US $) (cost)	Usual care + semaglutide (cost) (US $)	ICER ($)	ICER (€)
Deterministic analysis
0	20.44	20.91	119 900	415 446	625 863	575 793
1–99	19.24	19.91	124 235	408 710	425 492	391 453
100–399	17.78	18.68	127 220	395 346	296 505	272 785
≥400	14.18	15.52	128 442	355 107	168 666	155 172
Probabilistic sensitivity analysis
0	21.04	21.48	122 423	424 946	688 949	633 833
1–99	19.93	20.55	127 102	418 989	468 780	431 258
100–399	18.55	19.40	130 598	406 774	327 377	301 186
≥400	15.14	16.41	133 220	369 470	185 992	171 113

Note: This table reports the quality‐adjusted life years (QALY), total costs, and incremental cost‐effectiveness ratios (ICER) calculated for using semaglutide in addition to usual care in the 40‐year Markov model. The probabilistic sensitivity analysis was conducted using 1000 iterations and parameters used for this model are provided in Table S1. ICERs in euros (€) were calculated using a conversion rate of 1 USD = 0.92 EUR (as of October 2025).

Abbreviations: $, United States Dollar; €, Euro; CAC, coronary artery calcium; ICER, incremental cost effectiveness ratio; PSA, probabilistic sensitivity analysis; QALY, quality adjusted life years.

5. DISCUSSION

In a real‐world cohort of semaglutide eligible registry participants, the baseline CAC score was associated with higher MACE risk over the 10‐year study period. Additionally, the hypothetical use of semaglutide therapy among participants demonstrated incremental potential to reduce in MACE and HF risk across CAC groups with highest‐risk people observed to receive maximal possible benefit. Although ICERs decreased with increasing CAC burden, indicating that semaglutide becomes relatively more economically favourable in higher‐risk groups, ICERs remained above commonly cited US willingness‐to‐pay thresholds in the base case. These findings highlight the potential role of CAC in prioritizing therapy rather than establishing cost‐effectiveness under conventional thresholds.

Semaglutide and other glucagon‐like peptide 1 receptor agonists have revolutionized the therapeutic options for diabetes and obesity. While initially approved for treatment in people with type 2 diabetes, the proportion of US adults eligible for semaglutide therapy has exploded, with a recent study reporting that 129.2 million US adults can receive semaglutide therapy for weight management. ¹⁹ A poll conducted in 2024 of 1500 people reported that two in five have used semaglutide, and more than half of those reported that costs were untenable over the long‐term, despite insurance coverage ²⁰ and another study reported substantial inequity in semaglutide access. ²¹ Between 2022 and 2025, the Food and Drug Administration had issued a shortage notice for both Weigovy and Ozempic. ²² In such situations, pragmatic use of such drugs according to risk stratification may reduce access inequity and improve the overall health‐care costs. Therefore, our approach of using the CAC score as a risk‐stratification tool to model semaglutide use may have practical implications. Our study supports prior evidence from the MESA cohort that using CAC scores to adjudicate semaglutide therapy may be an efficient way to allocate limited healthcare resources. ²³ However, our study expanded on their analysis by applying it to a much larger real‐world cohort and investigating the actual cost benefit by lifetime modelling of real‐world scenarios. Importantly, while using different CAC cut‐offs (CAC > 300 vs. CAC > 400) to define the high‐risk group, our analysis actually reported that the NNT values may even be higher than those observed in the MESA cohort. Additionally, our modelling also demonstrated substantial potential benefit protection against incident HF. Given the strong causal relationship between obesity and heart failure with preserved ejection fraction (HFpEF) and the proven benefit of this drug in HFpEF patients, ²⁴ CAC could also be used to decide semaglutide therapy in people with undiagnosed HFpEF. Unfortunately, our analysis reported that, at current costs in the US, wider use of semaglutide is still prohibitively expensive. Prior studies from Portugal (ICER € 13 459/QALY) ²⁵ and the UK (ICER £14 827/QALY) ²⁶ have reported that semaglutide use for chronic obesity is cost‐effective in their healthcare systems. However, while model parameters may also influence ICER values, in this case, drug cost is the main predictor. A study reported that the launching cost for Ozempic in France and Germany was 143% and 572% lower than that in the US. ²⁷ Therefore, pragmatic treatment selection is likely the most efficient way to promote equitable healthcare access, given the current limited healthcare resources. Large prior evidence that cardiovascular risk is highest in the socially disadvantaged communities makes this approach even more pertinent. ²⁸ , ²⁹

5.1. Strength and limitations

Our study should be interpreted based on certain limitations. First, the CLARIFY registry is restricted to individuals without established cardiovascular disease, whereas SELECT enrolled participants with established cardiovascular disease. As a result, the baseline risk profile and absolute event rates in this analysis differ from those observed in SELECT. Our findings should therefore not be interpreted as directly generalizable to the SELECT trial population but rather as an evaluation of how CAC burden may inform risk‐based prioritization of semaglutide therapy in a broader primary prevention context. Second, outcome definitions differed from SELECT. While SELECT's primary endpoint included cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke, this analysis incorporated nonfatal myocardial infarction, nonfatal stroke, and all‐cause mortality due to data availability. These differences may influence the comparability of event rates and effect estimates across studies. Third, this was an observational analysis and subject to residual confounding. Fourth, semaglutide use was modelled rather than observed, relying on risk reductions from the SELECT trial and real‐world effects in individuals may vary, and finally, outcome was defined by the ICD codes and may be subject to misclassification. However, despite these shortcomings, our study provides novel evidence that CAC scores determine cardiovascular risk in a real‐world cohort and models the potential benefit of using risk stratification based on CAC scores to initiate semaglutide therapy.

6. CONCLUSION

In this real‐world analysis of obese adults without cardiovascular disease or diabetes, risk stratification using enrollment CAC scores correlated with observed adverse cardiovascular events. Our preliminary evidence suggests that using CAC scores for risk stratification may help guide semaglutide allocation in obese individuals and possibly reduce healthcare costs.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

FUNDING INFORMATION

None.

Supporting information

Data S1. Supporting Information.

Data S2. Supporting Information.

Ponnana SR, Zhang T, Sirasapalli SK, et al. Risk stratification using coronary artery calcium and potential benefit of semaglutide therapy: A cost‐effectiveness modelling study. Diabetes Obes Metab. 2026;28(4):3229‐3237. doi: 10.1111/dom.70515

DATA AVAILABILITY STATEMENT

Data for this study cannot be made available to researchers due to privacy reasons. However, a limited deidentified set can be requested from the corresponding author. Scripts used for all analyses in this study can be downloaded from the corresponding author's (SVD) Github account (https://www.github.com/svd09).