Background: The SELECT trial showed cardiovascular benefits of glucagon-like peptide-1 receptor agonists (GLP-1 RAs) in obese patients with cardiovascular disease; however, real-world data (RWD) on this benefit remain limited. This study used an artificial intelligence (AI)-generated algorithm and multimodal RWD to evaluate the impact of GLP-1 RAs on cardiovascular disease risk in a population of obese patients with and without preexisting cardiovascular disease. Methods: Using data from the Dandelion Health RWD library, an Emulated SELECT Cohort was created to include obese patients similar to those in the SELECT trial, but with and without preexisting cardiovascular disease. An AI algorithm developed by Pheiron that used 12-lead electrocardiograms (ECGs) as a predictive biomarker for the risk of major adverse cardiovascular events (MACE) was validated and used to derive MACE risk scores for the Emulated SELECT Cohort. These outcomes were compared over time between patients who used GLP-1 RAs and non-users using inverse-probability weighted linear regression models, adjusting for key covariates. Results: Out-of-sample validation showed high predictive accuracy of the AI algorithm, with ROC AUCs of 0.81 for myocardial infarction (MI) and 0.75 for stroke. Increased risk scores from the algorithm were correlated with higher MACE incidence in RWD. In the Emulated SELECT Cohort of 20,795 patients, GLP-1 RA use was associated with significant attenuation of MACE risk, with reductions observed in percentile risk score for MI (4%; p<0.001) and stroke (3.6%; p<0.001) per year of use. Differences in GLP-1 RA and non-GLP-1 users were evident as early as 1.7 years, with a 15-20% difference in absolute MACE risk scores between GLP-1 RA users and non-users observed by the end of the study. Conclusion: An AI algorithm using 12-lead ECGs accurately predicted MACE risk and could be used to model risk attenuation associated with GLP-1 RA use. Using this outcome, we find that GLP-1 RA use was associated with significant reductions in MI and stroke risk, in a broader population and within a shorter timeframe than the SELECT trial. These findings suggest potentially significant cardioprotective benefits of GLP-1 RAs in real-world settings and demonstrate a proof-of-concept for utilizing clinical AI to understand these benefits.

Shivaani Prakash is an employee of Dandelion Health AI, which provided the dataset for the study and the support for developing this manuscript. Jeffrey Coleman is an independent contractor who received funding from Dandelion Health AI to support the development of this manuscript.

This study did not receive any external funding.

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

All health record data utilized in this study are de-identified by expert determination under the HIPAA Privacy Rule before being made available to any users or researchers. In accordance with 45 C.F.R. Paragraph 46.101 Protection of Human Subjects, our study did not require Institutional Review Board approval because it used only de-identified medical data. All data used in this study are available to Dandelion Health AI and Dandelion Health AI gave ethical approval for this work.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

All data used in this study are available to Dandelion Health AI users.