Melatonin and cardiovascular risk: an NHANES reality check after recent heart failure concerns Authors: Baryalay Khan, Bruno Bezerra Lima Published: 2026 ABSTRACT Abstract BackgroundA large electronic health record study presented at the American Heart Association Scientific Sessions 2025 reported a heart failure safety signal associated with long-term melatonin use among adults with insomnia. In parallel, melatonin use has increased in the US general population, making counselling questions increasingly common in cardiovascular practice.Main argumentObservational associations between melatonin and cardiovascular outcomes are unusually vulnerable to confounding-by-indication, reverse causality, and exposure misclassification because melatonin uptake often reflects insomnia severity, depression or anxiety symptoms and healthcare utilisation.Evidence anchorIn a nationally representative NHANES (National Health and Nutrition Examination Survey) adult cohort (1999–2018), melatonin users differed markedly from non-users at baseline, particularly for sleep disturbance and depression. In updated survey-weighted, multiple-imputation models, recent melatonin use was not significantly associated with prevalent composite cardiovascular disease or heart failure. These findings provide a hypothesis-generating contextual counterpoint to recent alarm signals rather than causal evidence of benefit or harm.ConclusionAvailable observational data do not support strong claims of cardiovascular benefit, and they do not establish definitive cardiovascular harm or safety. Clinicians should document over-the-counter use, counsel pragmatically and prioritise evaluation and treatment of underlying sleep disorders while better evidence is developed.Keywords: Heart Failure, Risk Factors, Pharmacology, Epidemiology, Research Design Why this matters now Melatonin has long been perceived as a benign over-the-counter sleep aid. That perception was challenged by a large electronic health record study presented at the American Heart Association Scientific Sessions 2025.1 In that analysis of adults with a diagnosis of insomnia, long-term melatonin use recorded in the health record was associated with a higher incidence of heart failure over 5 years, alongside higher heart failure hospitalisation and all-cause mortality. Although the findings were not peer reviewed at the time of presentation and cannot establish causality, they received wide attention and created understandable clinical concern.2 The immediate question for cardiovascular clinicians is not whether melatonin has any biological activity, but whether the emerging observational signal should change counselling and prescribing behaviour. Answering that requires clarity on the population studied, the exposure definition and the degree to which melatonin use is a marker for underlying risk rather than a cause of risk. Why observational melatonin heart failure signals are hard to interpret Melatonin use is typically not random. Patients start melatonin because of insomnia symptoms, circadian disruption, stress, depression or anxiety symptoms and medication effects.3,5 These indications are themselves associated with cardiovascular risk and healthcare utilisation.3 6 7 This makes confounding-by-indication a dominant threat to inference. Reverse causality is also plausible. Early manifestations of declining health, including subclinical heart failure, can worsen sleep and trigger melatonin uptake. Insomnia symptoms themselves have been associated with incident heart failure in prospective cohort data.7 In this setting, melatonin can appear to precede heart failure when it is actually a response to emerging symptoms. Exposure misclassification is common because melatonin is frequently purchased over the counter and may not be consistently captured in electronic prescriptions, medication lists or population surveys.1 Even when use is reported, dose, timing, formulation, duration and persistence are often unavailable or inconsistently measured. In National Health and Nutrition Examination Survey (NHANES) specifically, exposure reflects self-reported use in the preceding 30 days and does not establish long-term use, cumulative dose or continuation after the survey examination. Independent analyses have shown substantial variability between labelled and measured melatonin content in supplements, which further complicates any dose response interpretation.8,10 Finally, adjustment choices can change the apparent signal. Variables such as sleep disturbance and depression strongly predict melatonin use. Treating them as standard confounders may reduce confounding, but it can also condition on the indication pathway and create overadjustment or collider bias. This is particularly relevant when the outcome of interest is heart failure, which is itself linked to sleep quality, depression and healthcare contact.3 6 7 A nationally representative reality check using NHANES To place recent alarm signals in context, we examined melatonin use and cardiovascular outcomes in the US NHANES from 1999 to 2018. NHANES provides a nationally representative sample with standardised baseline assessment. We defined exposure as any melatonin use in the past 30 days, recognising that this captures recent use rather than long-term therapy. Figure 1 shows the steady increase in 30-day melatonin use across survey cycles, consistent with previously reported national trends in melatonin supplement use; this increasing uptake makes melatonin counselling increasingly relevant in cardiovascular practice and underscores the need for clearer long-term safety evidence.² Prevalence of 30-day melatonin use by NHANES cycle. Survey-weighted prevalence of any melatonin use in the past 30 days among US adults aged 20 years and older, NHANES 1999–2018. The figure highlights increasing uptake over time, which makes melatonin counselling increasingly relevant in cardiovascular practice and underscores the need for clearer long-term safety evidence. NHANES, National Health and Nutrition Examination Survey. We analysed NHANES 1999–2018 adults aged 20 years and older with non-missing melatonin exposure. The primary NHANES endpoint was composite prevalent cardiovascular disease, defined as self-reported heart failure, coronary heart disease, myocardial infarction, angina or stroke, with heart failure examined as a secondary outcome. Survey-weighted logistic regression models used pooled Mobile Examination Center (MEC) examination weights and incorporated NHANES strata and primary sampling units to account for the complex survey design. Core adjusted models included age, sex, race and ethnicity, body mass index (BMI), smoking status, hypertension and diabetes. Depression and sleep disturbance were included in a clinical-burden sensitivity model because they are strongly related to the indication for melatonin use. Missing BMI, depression and sleep disturbance were addressed using multiple imputation, with survey design variables included in the imputation models. Because NHANES captures recent supplement exposure and prevalent cardiovascular outcomes, this analysis was intended as a hypothesis-generating contextual analysis rather than a causal test of melatonin safety or efficacy. Table 1 shows baseline characteristics by melatonin use status. The imbalance is substantial and clinically informative. Compared with non-users, melatonin users were older and were far more likely to report sleep disturbance (57.9% vs 20.5%) and depression (13.3% vs 5.5%). These differences support the view that melatonin use is strongly indication-driven. These imbalances have direct clinical implications: melatonin users represent a selected subgroup with greater sleep-related and psychiatric burden, older age and more cardiometabolic comorbidity, rather than an exchangeable comparison group. This reinforces that crude differences between users and non-users are vulnerable to confounding by indication and should not be interpreted causally. Table 1Baseline characteristics by melatonin use (past 30 days)CharacteristicNo melatoninMelatoninSample size, unweighted50 662267Age, years47.1 (17.0)51.4 (15.6)Female, %51.256.4Race/ethnicity, % Mexican American8.12.5 Other Hispanic5.73.1 Non-Hispanic white68.089.3 Non-Hispanic black11.32.2 Other/multiracial7.03.0Body mass index, kg/m²28.7 (6.7)29.7 (6.7)Smoking status, % Never53.749.3 Former28.337.5 Current18.013.2Hypertension, %30.640.6Diabetes, %10.813.2Depression, %5.513.3Sleep disturbance, %20.557.9 Values are survey-weighted unless stated otherwise. Continuous variables are mean (SD). Baseline characteristics by melatonin use (past 30 days) Values are survey-weighted unless stated otherwise. Continuous variables are mean (SD). In the updated survey-weighted, multiple-imputation analysis, melatonin use was not significantly associated with prevalent composite cardiovascular disease in the core model adjusted for age, sex, race and ethnicity, BMI, smoking status, hypertension and diabetes (adjusted OR 0.90, 95% CI 0.58 to 1.40). In a clinical-burden sensitivity model additionally adjusting for depression and sleep disturbance, the estimate was 0.72 (95% CI 0.46 to 1.11). Heart failure was examined as a secondary exploratory outcome. Corresponding estimates were 0.88 (95% CI 0.43 to 1.80) in the core model and 0.67 (95% CI0.33 to 1.38) in the clinical-burden model. These estimates should not be interpreted as evidence that melatonin is cardioprotective, nor do they establish long-term safety. Rather, they indicate that in this exploratory NHANES analysis, recent melatonin use was not associated with statistically significant higher odds of prevalent composite cardiovascular disease or heart failure after adjustment. The direction and imprecision of estimates should be interpreted in the context of indication-driven exposure, residual and structural confounding, selection and limited numbers of melatonin users with cardiovascular outcomes. Reconciling an insomnia cohort signal with NHANES The TriNetX insomnia cohort and NHANES answer different questions. The American Heart Association 2025 analysis was restricted to patients with clinically recognised insomnia and evaluated long-term melatonin use recorded in the health record.1 In contrast, NHANES captures self-reported recent supplement use in the general population. These differences change the baseline risk profile, the severity of insomnia, the likelihood of psychiatric comorbidity and the completeness of exposure capture. Additional sources of discrepancy are also important. In electronic health record (EHR)-based insomnia cohorts, melatonin exposure, insomnia severity, psychiatric comorbidity, concurrent sleep therapies and cardiovascular risk may change over time. These time-varying factors can influence both continued melatonin use and later heart failure diagnosis. EHR medication lists may also incompletely capture over-the-counter use, discontinuation, dose and formulation, while outcome ascertainment may be influenced by healthcare utilisation and coding practices. In contrast, NHANES captures self-reported 30-day supplement use at baseline in a general population sample and cannot determine long-term persistence or changes in exposure after the survey examination. As a result, an apparent heart failure association in an insomnia cohort could reflect insomnia severity, unmeasured psychiatric disease, concurrent unrecorded sleep therapies or healthcare utilisation patterns. Conversely, a null or non-harm signal in NHANES does not prove long-term safety for high-dose, long-duration use in severe insomnia. Therefore, the contrast between the EHR insomnia cohort and NHANES should be interpreted as evidence of methodological heterogeneity rather than as a direct contradiction between studies. Taken together, the evidence supports caution in causal interpretation and argues against strong claims of either harm or benefit based solely on observational associations. Practical implications for cardiovascular clinicians A pragmatic clinical stance is appropriate. Current observational evidence does not justify recommending melatonin for cardiovascular prevention. At the same time, exploratory NHANES data should not be interpreted as establishing long-term cardiovascular safety or directly refuting EHR-based findings. For patients using low-dose, short-term melatonin, the current evidence supports cautious, pragmatic counselling rather than alarm or reassurance based on causal claims. However, evidence supporting melatonin for chronic insomnia in adults is mixed and guideline recommendations do not support routine use.4 5 Clinicians should document over-the-counter use explicitly, particularly in older adults and patients with polypharmacy. Counselling should prioritise evaluation for treatable sleep disorders, behavioural sleep interventions and careful review of concurrent sedatives. Given product variability, patients should be advised to use reputable products and the lowest effective dose for the shortest duration.8,10 Randomised trial data suggest adverse events are generally uncommon and mild, but long-term safety evidence remains limited.11 What evidence should come next Resolving the heart failure question will require designs that match real-world use. Pragmatic trials in populations with persistent insomnia symptoms, along with target trial emulation studies using longitudinal medication capture, could reduce confounding-by-indication and address time-varying confounding. Future observational work should measure dose, timing, formulation and persistence, and should separately evaluate short-term use for circadian indications vs long-term use for chronic insomnia. Outcomes should be clinically aligned, including incident heart failure, atrial fibrillation, myocardial infarction, stroke and patient-centred sleep outcomes. In the interim, both clinicians and patients should interpret emerging observational signals with caution, particularly when the exposure is heavily indication-driven. Conclusion Melatonin use is increasing and the cardiovascular community is now confronting high-profile observational heart failure safety concerns. A nationally representative NHANES analysis provides a hypothesis-generating contextual counterpoint to recently reported EHR-based associations between long-term melatonin use and higher heart failure risk. Recent melatonin use was not associated with statistically significant higher odds of prevalent composite cardiovascular disease or heart failure in updated survey-weighted analyses. These findings should not be interpreted as causal evidence of benefit or harm, but they underscore the need for better longitudinal studies before strong safety conclusions are drawn. Footnotes Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Patient consent for publication: Not applicable. Ethics approval: This study used publicly available, de-identified data from the National Health and Nutrition Examination Survey (NHANES) and linked mortality files. No separate institutional ethical approval was required for this secondary analysis. The underlying NHANES protocols were approved by the National Center for Health Statistics Research Ethics Review Board, and all participants provided written informed consent. Provenance and peer review: Not commissioned; externally peer reviewed. Data availability free text: The data used in this study are publicly available from the National Health and Nutrition Examination Survey (NHANES), including the linked mortality files made available through the National Center for Health Statistics. References