Melatonin affects more than sleep. It can change how the body handles insulin and blood sugar, which belongs in any serious discussion of its potential cardiovascular effects.
The late-2025 headlines linking melatonin to heart failure deserve scrutiny. My concern is broader than one headline: a sleep supplement should provide a meaningful benefit, and its metabolic effects should not be ignored.
What the heart-failure study found:
A study presented at the American Heart Association’s 2025 Scientific Sessions compared 130,828 adults with insomnia. Over five years, heart failure was recorded in approximately 4.6% of those with documented melatonin use for at least a year and 2.7% of the comparison group. That is an observed difference of about 19 diagnoses per 1,000 people. The reported hazard ratio was 1.89, the source of the “nearly 90% higher risk” headline. A hazard ratio and the ratio of those percentages are different measures. Read the AHA study report.
Both groups had insomnia. The unresolved issue is whether the groups were otherwise comparable: over-the-counter melatonin can be absent from records, prescribing rules vary by country, and insomnia severity and some psychiatric conditions were not captured. “No melatonin recorded” is not reliable proof of no use.
This was a preliminary conference abstract based on medical records. A randomized trial with verified exposure and comparable insomnia severity would better test whether melatonin causes heart failure. The association warrants investigation, but it does not answer that causal question.
The glucose findings deserve attention:
We do have controlled human experiments showing adverse metabolic effects under specific conditions.
In a randomized, placebo-controlled crossover trial, 17 men with type 2 diabetes took 10 mg of melatonin one hour before bed for three months and placebo for another three months. Researchers used a glucose clamp, a direct experimental measure of insulin sensitivity. Insulin sensitivity fell by approximately 12% during melatonin treatment.
That is a measured adverse physiological effect in a controlled experiment. The trial was small and tested one dose in men with diabetes, so it cannot tell us what happens with every dose or in every person. It does establish a reason to take the metabolic question seriously.
These glucose findings add biological plausibility to concerns about melatonin. They do not establish that impaired glucose regulation caused the heart-failure association.
Why genetics and timing matter:
MTNR1B encodes a melatonin receptor. In research published in 2016, a common diabetes-risk variant was linked to increased receptor expression in human pancreatic islets. In the human intervention, melatonin reduced insulin secretion and raised glucose more in risk-variant carriers. This helps explain why the same supplement may affect people differently.
A 2026 randomized crossover trial studied 21 healthy participants, including 10 risk-allele carriers. A 5 mg dose worsened glucose tolerance and reduced modelled first-phase beta-cell responsiveness, the early insulin response to glucose, by approximately 40% in carriers. Corresponding statistically significant effects were not detected in noncarriers. This was an acute daytime laboratory experiment, not a long-term bedtime safety trial.
Meal timing adds context. A randomized crossover study of 845 adults used glucose drinks to simulate earlier and later dinners. Glucose tolerance was worse near bedtime, when natural melatonin was higher, especially in risk-allele carriers. It tested timing rather than melatonin supplements, but reinforces why the timing of food and melatonin exposure matters.
The evidence pointing the other way:
A balanced review must include favorable findings. A 2026 meta-analysis of 30 randomized trials found improvements in fasting glucose, insulin and HOMA-IR, an estimate of insulin resistance. The average fasting-glucose reduction was modest, about 2.7 mg/dL, and results varied considerably across trials.
A separate 2026 meta-analysis found improvements in insulin-resistance markers but no significant fasting-glucose reduction. Fasting markers, clamp measurements and responses to a glucose challenge measure different aspects of metabolism. Differences in participants, dose, timing and duration may also contribute to the mixed findings.
The MeHR randomized trial studied people who already had heart failure with reduced ejection fraction. Ninety-two patients were recruited and 85 completed 24 weeks of 10 mg nightly melatonin or placebo. Melatonin improved NT-proBNP, a marker of cardiac stress, and quality of life. This trial studied existing heart failure, not new diagnoses, and was too small to establish an effect on survival.
A 2026 NHANES analysis also found no significant association between recent melatonin use and existing cardiovascular disease or heart failure after adjustment. Its different population and exposure window do not resolve the long-term insomnia study, but they add to the uncertainty around cardiovascular outcomes.
My approach:
If melatonin meaningfully improves sleep and quality of life, it may have a place. That benefit should be weighed against the dose, duration, timing and the person’s metabolic health.
If it is not providing a clear benefit, I do not see a good reason to keep taking it out of habit. There is enough evidence of potential metabolic harm to question routine use, particularly at higher doses, while long-term cardiovascular effects remain unresolved. Review ongoing use with your clinician, especially if you have diabetes or concerns about blood sugar.
For our current product-selection approach and sleep-related options, see the Toward Health Supplement Guide, including “Sleep, recovery and the nervous system.” The guide supports an individualized conversation; it is not a recommendation to take every listed supplement.
Dr. Tro Kalayjian
