Heart-related problems reported with tirzepatide: real-world data on heart failure, irregular beats, and blood flow blockages

The global prevalence of obesity and type 2 diabetes mellitus (T2DM) has reached epidemic proportions, driving an urgent need for effective therapeutic interventions that address both metabolic dysregulation and associated cardiovascular risks (Shmygol et al., 2025; Campos et al., 2025). Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have transformed the management of these conditions, demonstrating potent glucose-lowering and weight-loss effects alongside established cardiovascular benefits (Solini et al., 2023; Sheth et al., 2025). Building on this success, tirzepatide, a novel first-in-class dual glucose-dependent insulinotropic polypeptide (GIP) and GLP-1 receptor agonist, has emerged as a transformative agent. By synergistically activating both incretin receptors, tirzepatide has demonstrated superior efficacy in glycemic control and body weight reduction compared to selective GLP-1RAs in the landmark SURPASS (Study of Tirzepatide in Participants With Type 2 Diabetes) and SURMOUNT (Study of Tirzepatide in Participants With Obesity) clinical trial programs (Samms et al., 2022; Novikoff and Muller, 2023).

Despite its promising metabolic profile, the cardiovascular safety of tirzepatide remains a subject of intense clinical interest and ongoing investigation. While GLP-1RAs are known to reduce the risk of major adverse cardiovascular events (MACE) in high-risk populations (Badve et al., 2025), they are also associated with a consistent increase in resting heart rate, raising theoretical concerns about potential arrhythmogenic risks or deleterious hemodynamic effects over the long term (Greco et al., 2022). The initial data from the Phase three trial of tirzepatide showed a dose-dependent increase in heart rate (Yang et al., 2024). Although a pre-specified meta-analysis of the SURPASS trials suggested no increase in MACE risk (Sattar et al., 2022), these trials were not primarily powered to adjudicate specific cardiovascular outcomes such as heart failure or atrial fibrillation, and patients with unstable cardiovascular conditions were often excluded.

Furthermore, recent evidence has sparked a paradigm shift regarding the potential of incretin-based therapies in heart failure management. The STEP-HFpEF trial (Semaglutide Treatment Effect in People with obesity and HFpEF) recently demonstrated that semaglutide significantly reduced symptoms and physical limitations in patients with heart failure with preserved ejection fraction (HFpEF) and obesity, suggesting a direct cardioprotective role beyond weight loss (Schou et al., 2024; Kosiborod et al., 2023). However, whether tirzepatide, with its additional GIP component, confers similar or distinct cardiovascular benefits—or risks—in a broader, unselected real-world population remains largely unknown. Post-marketing surveillance is crucial to bridge the gap between controlled clinical trials and real-world clinical practice, where patient heterogeneity regarding age, comorbidities, and polypharmacy is substantially higher (Wang M. et al., 2025; Hunter Gibble et al., 2025).

Therefore, this study aimed to characterize the real-world cardiovascular adverse event reporting profile of tirzepatide using the FAERS database from 2022 to 2025. Specifically, we sought to: (1) assess the reporting frequency of major cardiovascular events, including heart failure, atrial fibrillation, myocardial infarction, angina, myocardial ischemia, and tachycardia, relative to a concurrent full-database comparator; (2) evaluate whether these reporting patterns are modified by baseline factors such as body weight, age, and sex; and (3) analyze longitudinal trends to determine the stability of any observed associations over time.

We identified cases associated with tirzepatide using a text-mining approach. The DRUGNAME column was searched using a case-insensitive regular expression query to capture both generic and brand names, including: “tirzepatide”, “Mounjaro”, “Zepbound”, and the research code “LY3298176.”

To minimize confounding bias and focus on drug-induced events, we restricted the analysis to reports where tirzepatide was listed as the “Primary Suspect” (PS) drug (Figure 1). Concomitant or interacting drugs were excluded from the exposure definition.

Disproportionality analysis, utilizing the full FAERS database as the background comparator, characterized the reporting profile of tirzepatide regarding specific cardiovascular outcomes (Table 2).

In the full dataset, tirzepatide was associated with lower reporting odds across all assessed cardiovascular events, including heart failure (ROR 0.08 [95% CI 0.07–0.11]), atrial fibrillation (ROR 0.29 [0.24–0.35]), tachycardia (ROR 0.62 [0.57–0.67]), acute myocardial infarction (ROR 0.32 [0.27–0.37]), angina (ROR 0.14 [0.08–0.23]), and myocardial ischemia (ROR 0.12 [0.04–0.36]).

In the sensitivity analyses restricted to healthcare professional (HCP) reports, the reporting profile demonstrated notable divergences. Tirzepatide remained associated with persistently lower reporting odds for heart failure (ROR 0.18 [0.07–0.43]). Regarding ischemic outcomes, no significant disproportionate reporting was detected for acute myocardial infarction (ROR 0.45 [0.19–1.08]). Strikingly, there were zero reported cases of either angina or myocardial ischemia in the HCP cohort, rendering the respective RORs not calculable (NC) but heavily underscoring the lack of any ischemic safety signal in clinical practice. In contrast to the findings in the full dataset, no significant association was detected for atrial fibrillation (ROR 1.02 [0.58–1.80]) or tachycardia (ROR 1.01 [0.68–1.51]) after accounting for reporter bias in HCP reports.

This study, analyzing over 100,000 pharmacovigilance reports from the FDA FAERS database, provides a comprehensive real-world assessment of the cardiovascular adverse event reporting profile of tirzepatide. A critical methodological consideration in our analysis is the overwhelming proportion of consumer-submitted reports (92.3%). Because consumer reports are highly susceptible to subjective interpretation—where physiological palpitations may be inaccurately coded as “tachycardia,” and complex conditions like heart failure may be underreported due to a lack of formal medical diagnosis—our primary clinical inferences are drawn from sensitivity analyses restricted to healthcare professional (HCP) reports. Utilizing a concurrent full-database comparator, our HCP-restricted analysis reveals a reporting profile characterized by persistently lower reporting odds for heart failure (ROR 0.18, 95% CI: 0.07–0.43). Concurrently, the signals for atrial fibrillation (ROR 1.02, 95% CI: 0.58–1.80) and tachycardia (ROR 1.01, 95% CI: 0.68–1.51) shifted to the null in the HCP cohort. Similarly, ischemic events including acute myocardial infarction shifted to neutrality (ROR 0.45, 95% CI: 0.19–1.08), while remarkably, angina and myocardial ischemia recorded zero cases in the HCP subset. These findings suggest no disproportionate reporting of these arrhythmic and ischemic events when restricted to medically verified cases.

It is imperative to emphasize that FAERS disproportionality analyses reflect reporting patterns rather than true reductions or elevations in clinical risk (Donati et al., 2025; Irizarry et al., 2025). Therefore, a Reporting Odds Ratio (ROR) of less than one must not be misconstrued as clinical “cardioprotection.” Lower RORs in spontaneous reporting systems frequently result from competing adverse event reporting. As demonstrated in our landscape analysis (Figure 3), the tirzepatide reporting profile is heavily dominated by gastrointestinal adverse events and administration-related errors, which mathematically depress the proportional reporting of cardiovascular events (Garvey et al., 2023; Rubino et al., 2025). Furthermore, these lower reporting odds may be significantly influenced by confounding by indication and channeling bias, wherein patients with severe pre-existing cardiovascular disease or advanced heart failure might be preferentially prescribed alternative therapies or closely monitored, altering the spontaneous reporting dynamic (Sendaydiego et al., 2024; Grubic et al., 2025).

Nevertheless, the persistently lower reporting odds for heart failure, which remained robust in the HCP-restricted analysis, is a notable finding that aligns with recent paradigm-shifting clinical trials. For example, the STEP-HFpEF trial demonstrated that semaglutide significantly improved outcomes in patients with heart failure with preserved ejection fraction (HFpEF) (Kosiborod et al., 2023; Kosiborod et al., 2024). Our real-world data suggest that the dual GIP/GLP-1 agonist tirzepatide does not exhibit a safety signal for heart failure, which is consistent with the hypothesized underlying mechanisms that extend beyond weight reduction. Specifically, the additional activation of GIP receptors by tirzepatide may confer unique advantages in adipose tissue and the vasculature, hypothesized to enhance the “buffering capacity” of white adipose tissue, preventing myocardial lipotoxicity (Campbell et al., 2022; Mestres-Arenas et al., 2026). This concept is further reinforced by our findings regarding ischemic outcomes. The complete absence of angina and myocardial ischemia reports in the HCP-restricted cohort, alongside neutral signals for acute myocardial infarction, strongly complements the proposed anti-atherosclerotic potential of GLP-1/GIP dual agonism to improve endothelial function and plaque stability (Pujadas et al., 2022).

Regarding atrial fibrillation (AF) and tachycardia, the inverse associations observed in the full dataset were not replicated in HCP-restricted analyses. This divergence strongly suggests that the initial low reporting was an artifact of consumer reporting patterns rather than a true clinical absence of the events. Furthermore, our weight-stratified analyses revealed elevated reporting odds for tachycardia in both the lower (<95 kg: ROR 1.91) and higher (≥95 kg: ROR 1.57) weight groups. However, these findings must be interpreted with extreme caution. Body weight data were missing in 94.3% of the total cohort, meaning these stratified analyses were confined to a highly selected subset of only 5,881 reports. Patients with meticulously documented weight in spontaneous reporting systems often represent cases undergoing intensive clinical monitoring. In such meticulously documented sub-cohorts, expected pharmacological effects—such as the known dose-dependent chronotropic effect of GLP-1 receptor agonists (increases of two to four beats per minute documented in the SURPASS and SURMOUNT trials)—are more likely to be “unmasked” and captured (Yang et al., 2024; Ayesh et al., 2024). We strictly refrain from drawing mechanistic conclusions regarding the weight-related heterogeneity of atrial fibrillation based on this incomplete subgroup reporting, as the unclear denominators and extensive missing data preclude reliable biological inferences.

Although elevated heart rate is a traditional cardiovascular risk factor, our longitudinal trend analysis (Figure 4) indicates this signal is non-progressive. While a visibly higher reporting rate of tachycardia was observed in 2022, this initial peak is characteristic of the “Weber effect”—a well-documented pharmacovigilance phenomenon where adverse event reporting surges in the first year following a drug’s approval due to heightened clinical vigilance and novelty (Carriot et al., 2021; Gérard et al., 2022).

This study has major limitations inherent to spontaneous reporting systems. First, disproportionality analysis measures statistical associations in reporting frequencies and does not establish definitive causality or quantify absolute risk. Second, the lack of a denominator (total exposed population) prevents the calculation of true incidence rates. Third, the absence of systematic data on critical clinical confounders severely limits our ability to perform multivariable adjustments. Because FAERS is a spontaneous reporting system heavily driven by consumer reports (92.3%), documentation of past medical history and chronic metabolic or cardiovascular comorbidities is overwhelmingly missing or inconsistently coded. Furthermore, FAERS lacks an “untreated” control cohort, relying instead on a background comparator of patients experiencing adverse events from other drugs. Consequently, we could not adjust for baseline cardiovascular risk factors, diabetes duration, smoking status, or concomitant medications. Fourth, although our study period (2022 Q2–2025 Q3) represents the entire post-marketing period currently available in FAERS for tirzepatide, this timeframe is relatively short. Longer observation periods would provide more stable safety signals and allow for the detection of late-emerging adverse events. Therefore, our findings should be interpreted purely as hypothesis-generating reporting patterns requiring validation in prospective, controlled cardiovascular outcome trials or large-scale electronic health record (EHR) studies.

In this large-scale FAERS analysis using a concurrent full-database comparator, tirzepatide was associated with persistently lower reporting odds for heart failure across all sensitivity strata, including HCP-restricted analyses. No association was detected for atrial fibrillation, tachycardia, or ischemic events (acute myocardial infarction, angina, and myocardial ischemia) after accounting for reporter bias. Elevated reporting odds for tachycardia observed in weight-stratified analyses were confined to a small subset with documented weight (5.7%) and were not replicated in HCP reports. These findings reflect database reporting patterns influenced by channeling bias and competing adverse events, and do not prove clinical cardioprotection. This real-world evidence complements the safety database of tirzepatide and generates hypotheses for ongoing cardiovascular outcome trials.

The author(s) declared that financial support was received for this work and/or its publication. This research was supported by Joint Funds for the Innovation of Science and Technology, Fujian province (Grant number: 2021Y9035).

Publicly available datasets were analyzed in this study. This data can be found here: FDA Adverse Event Reporting System (FAERS) Public Dashboard and Quarterly Data Extract Files (https://fis.fda.gov/extensions/FPD-QDE-FAERS/FPD-QDE-FAERS.html↗). Further inquiries regarding the processed data can be directed to the corresponding authors.

Ethical approval was not required for the studies involving humans because Not Applicable (Publicly available FAERS database). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants and legal guardians/next of kin in accordance with the national legislation and institutional requirements.

DC: Conceptualization, Data curation, Formal Analysis, Methodology, Software, Writing – original draft. ZW: Conceptualization, Methodology, Writing – original draft. ZX: Investigation, Validation, Visualization, Writing – original draft. YC: Investigation, Validation, Visualization, Writing – original draft. SL: Conceptualization, Project administration, Supervision, Writing – review and editing. SX: Conceptualization, Project administration, Supervision, Writing – review and editing.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The author(s) declared that generative AI was used in the creation of this manuscript. During the writing process, artificial intelligence technologies (such as Gemini and related technologies from Google) were employed to enhance the readability and language quality of the manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.