Atrial fibrillation, the most common sustained heart rhythm disorder worldwide, does not strike at random. For the millions of people living with coronary artery disease, the medications prescribed to protect their hearts may themselves shape the odds of developing this arrhythmia. That is the central conclusion of a new retrospective cohort study published in Biomedical Engineering Letters, in which researchers in South Korea analyzed prescription and outcome data from 157,750 patients with coronary artery disease and related comorbidities and found that the risk of new-onset atrial fibrillation varies substantially depending on which cardiovascular drugs a patient receives.
The research team, led by Taehyun Kim and Gayeon Ryu, who contributed equally, together with Ye Eun Choi and corresponding author Hangsik Shin of the University of Ulsan College of Medicine and Asan Medical Center in Seoul, set out to fill a gap in the evidence base. Although it has long been recognized that cardiovascular medications can modulate the risk of incident atrial fibrillation, comprehensive medication-specific data in the coronary artery disease population have remained limited. Because atrial fibrillation prevalence, incidence, and mortality have been rising for decades—as documented in landmark analyses such as the Framingham Heart Study’s 50-year trend review and the Global Burden of Disease project—understanding how routine prescribing decisions influence arrhythmia risk carries significant public health weight.
To identify which drugs deserved scrutiny, the investigators first performed a frequency analysis of prescriptions across the entire cohort of 157,750 patients, isolating thirty CAD-predominant medications spanning nine pharmacological classes. These included statin-based antilipidemic agents, calcium channel blockers, beta-blockers, diuretics, anticoagulants, and vasoactive drugs routinely deployed in the management of chronic and acute coronary syndromes. The analytical cohort was then narrowed to the 99,370 patients who had received at least one of these medications, and the researchers evaluated the association between each drug and incident atrial fibrillation using extended Cox proportional hazards models with time-varying covariates—a statistical framework, originally formalized by David Cox in 1972, that allows the effect of an exposure to change over the follow-up period and reduces the bias that arises when treatment status is treated as fixed.
The results were striking in their heterogeneity. Of the thirty medications examined, six showed statistically significant associations with atrial fibrillation risk in multivariable models. Five were linked to elevated risk and one to reduced risk. The cholesterol-lowering combination ezetimibe/rosuvastatin emerged as the sole protective agent, associated with a lower likelihood of developing the arrhythmia. In contrast, the calcium channel blocker diltiazem hydrochloride, the loop diuretic torasemide, the anticoagulant heparin sodium, the beta-blocker bisoprolol fumarate, and the vasopressor norepinephrine were each associated with increased risk. Among all the drugs studied, torasemide displayed the strongest positive association with new-onset atrial fibrillation.
The mechanistic story behind these findings is as varied as the drugs themselves. Torasemide and other diuretics can provoke hypokalemia—depletion of serum potassium—which is a well-established trigger of atrial arrhythmias. The Rotterdam Study and subsequent work have shown that low serum potassium levels predict atrial fibrillation risk, and electrophysiological experiments demonstrate that electrolyte disturbances differentially alter the electrical behavior of the sinoatrial node and the pulmonary veins, creating substrates for the re-entrant activity that sustains fibrillation. Diltiazem, paradoxically, is a mainstay for rate control in patients who already have atrial fibrillation, yet its negative inotropic and vasodilatory effects, along with the clinical circumstances prompting its prescription, may mark or promote arrhythmogenic conditions in CAD patients who have not yet developed the disorder. Norepinephrine’s association is perhaps the most biologically plausible of all: adrenergic stimulation is a classical driver of triggered activity, and animal models of ischemia and reperfusion have shown that norepinephrine directly promotes arrhythmic events, likely through calcium handling disturbances in atrial myocytes. Bisoprolol’s signal is more counterintuitive, since beta-blockers are widely used to suppress arrhythmias, but confounding by indication—the possibility that sicker patients receive these drugs—remains an ever-present caveat in observational pharmacoepidemiology.
Ezetimibe/rosuvastatin’s apparent protective effect aligns with a growing literature on the antiarrhythmic potential of lipid-lowering therapy. Earlier work in patients with coronary artery disease found that statins protect against atrial fibrillation, and mechanistic studies have shown that the antiarrhythmic effect may stem from reductions in atrial oxidative stress, since the sources of reactive oxygen species in the atria shift as fibrillation progresses. A meta-analysis of randomized controlled trials similarly concluded that statin therapy exerts an antiarrhythmic effect, and the new findings suggest that combining rosuvastatin with the cholesterol-absorption inhibitor ezetimibe may preserve or enhance this benefit in the CAD population.
At the level of drug classes, the study’s pattern was coherent: antilipidemic agents were collectively associated with lower atrial fibrillation risk, whereas certain diuretics, antianginal agents, antihypertensives, and vasoactive agents were associated with higher risk. This class-level dichotomy suggests that the lipid-lowering axis of CAD management may be doing double duty—protecting against atherosclerotic events while simultaneously damping arrhythmogenic processes such as inflammation, oxidative injury, and atrial structural remodeling. Conversely, drugs that perturb electrolyte balance, hemodynamics, or autonomic tone may tip vulnerable atria toward fibrillation.
The authors emphasize that these results should inform individualized pharmacotherapy rather than provoke alarm. Coronary artery disease and atrial fibrillation share a bidirectional relationship: genetic analyses using Mendelian randomization suggest each condition predisposes to the other, and studies of myocardial infarction highlight the importance of atrial ischemia in precipitating arrhythmias. Within this shared substrate, the choice of concomitant medication may meaningfully shift the trajectory. When two therapeutic options are otherwise equivalent, the new data argue that clinicians managing CAD patients should weigh atrial fibrillation risk among the selection criteria—an approach that could translate into meaningful prevention at the population scale, given the enormous number of patients on these therapies.
Methodologically, the study leveraged Korea’s national health information infrastructure, a resource that has enabled numerous large-scale cardiovascular investigations, and the extended Cox models with time-varying covariates addressed one of the classic pitfalls of medication-outcome research: the fact that patients start, stop, and switch drugs over time. By allowing exposure status to evolve during follow-up, the analysis better reflects real-world prescribing dynamics than static ever-versus-never comparisons. The work was approved by the Institutional Review Board of Asan Medical Center with informed consent waived owing to the retrospective, de-identified design, and was supported by the National Research Foundation of Korea and the Asan Institute for Life Sciences.
The investigators are candid about the limits inherent to retrospective cohort analysis. Association is not causation, and confounding by indication looms large for drugs such as norepinephrine, which is administered to hemodynamically unstable patients in whom atrial fibrillation risk is already elevated for reasons unrelated to the drug itself. Residual confounding by disease severity, unmeasured electrolyte levels, and differences in monitoring intensity—sicker patients receive more electrocardiograms, and thus more opportunities for arrhythmia detection—could all contribute to the observed signals. The data underlying the study are not publicly available because of institutional and ethical restrictions, though requests for de-identified data may be considered subject to review board approval.
Even with these caveats, the study represents one of the most comprehensive medication-level mappings of atrial fibrillation risk in the coronary artery disease population to date, and its conclusions arrive at a moment of rising global atrial fibrillation burden and escalating hospital care costs. As the authors note, the demonstrated heterogeneity of AF risk across CAD medications carries direct implications for personalized therapy and arrhythmia prevention. The next step—testing whether prospectively selecting lower-risk alternatives among therapeutically comparable drugs actually reduces incident atrial fibrillation—will require randomized or quasi-experimental designs. Until then, the message for cardiologists is a sober reminder that in cardiovascular medicine, even the drugs prescribed to protect the heart can, under some circumstances, reshape its rhythm.
Subject of Research: Medication-specific risk of incident atrial fibrillation in patients with coronary artery disease
Subject of Research: Technology and Engineering
Article Title: Medication-specific risk of atrial fibrillation in patients with coronary artery disease
Article References: Kim, T., Ryu, G., Choi, Y. E., & Shin, H. (2026). Medication-specific risk of atrial fibrillation in patients with coronary artery disease. Biomedical Engineering Letters. https://doi.org/10.1007/s13534-026-00589-6
Image Credits: AI Generated
DOI: 10.1007/s13534-026-00589-6
Keywords: atrial fibrillation risk, coronary artery disease, cardiovascular medications, ezetimibe/rosuvastatin, torasemide, Cox proportional hazards model, personalized therapy, survival analysis, drug-induced arrhythmia, pharmacotherapy
Cite Scienmag News
APA MLA Chicago
Phoebe Ingram. (September 9, 2026). Drug-specific atrial fibrillation risk seen in coronary artery disease patients. Scienmag. https://scienmag.com/drug-specific-atrial-fibrillation-risk-seen-in-coronary-artery-disease-patients/
Phoebe Ingram. “Drug-specific atrial fibrillation risk seen in coronary artery disease patients.” Scienmag, 9 September 2026, https://scienmag.com/drug-specific-atrial-fibrillation-risk-seen-in-coronary-artery-disease-patients/. Accessed 9 September 2026.
Phoebe Ingram. “Drug-specific atrial fibrillation risk seen in coronary artery disease patients.” Scienmag. September 9, 2026. https://scienmag.com/drug-specific-atrial-fibrillation-risk-seen-in-coronary-artery-disease-patients/
Copy citation Download RIS
Tags: analysis of drug-specific atrial fibrillation riskarrhythmia risk analysisatrial fibrillation in heart disease patientsatrial fibrillation riskatrial fibrillation risk in coronary artery disease patientscardiovascular drug impact on arrhythmiacardiovascular medication safety in coronary artery diseasecoronary artery diseasecoronary artery disease treatment outcomesdrug-specific atrial fibrillation risk factorsepidemiology of atrial fibrillation in heart disease patientsimpact of cardiovascular drugs on atrial fibrillation developmentimpact of cardiovascular medications on atrial fibrillationlong-term heart rhythm disorder trendslong-term trends in atrial fibrillation incidencemedication modulation of arrhythmia riskmedication-related atrial fibrillationmedication-specific atrial fibrillation risk factorsprescription data and heart rhythm disordersprescription patterns and atrial fibrillation outcomesretrospective cohort heart studyretrospective cohort study on atrial fibrillation and coronary artery disease
