Cardiovascular safety of melatonin: insights from nationwide Danish registries
D M Christensen, M Blenkov, M Pilacvi, H Christensen, C H Garred, C Sindet-Pedersen, A Holt, J Strange, C Friis, L Kober, C Torp-Pedersen, M SchouAbstract
Background
Melatonin is commonly used for the treatment of insomnia. It is considered harmless despite a lack of robust safety studies. Recent data raises concern over a potential risk of heart failure (HF) associated with long-term use(1). We sought to assess the cardiovascular safety associated with melatonin treatment using the nationwide Danish registries.
Methods
We identified all new users of melatonin from 2015-2023 aged >18 years without prevalent cardiovascular disease (CVD) at baseline. They were matched in a 1:1 ratio on age, date, and sex with non-user controls without prevalent CVD. Time zero was the date of first melatonin prescription and the corresponding matched date for controls. The outcomes were HF, acute coronary syndrome (ACS), arrythmia, and ischemic stroke (IS). Cause-specific Cox regression models accounting for the competing risk of death adjusted for comorbidities (including diabetes, mental illness) and concomitant pharmacotherapy (including antidepressants, opioids) were used to calculate 5-year standardized risks, risk differences, and risk ratios (sRR) for melatonin vs. controls overall and in subgroups. In addition, we performed active comparator (zopiclone and zolpidem, Z-drugs) and landmark cumulative daily defined dose (cDDD) sensitivity analyses. Importantly, in Denmark, melatonin and other hypnotics require a prescription and are not sold over the counter.
Results
In the main analyses, 273,167 new users of melatonin were matched with the same number of controls (median age 45 years, 62% female). Melatonin users had more comorbidities, including cancer (4.7% vs. 3.0%) and COPD (1.3% vs. 0.8%), and took more antidepressants (28.4% vs. 6.9%), opioids (13.5% vs. 5.6%), and benzodiazepines (8.1% vs 1.8%) at baseline. They also had a higher prevalence of risk factors for CVD (diabetes, 5.2% vs. 4.1%; hypertension, 8.6% vs. 7.6%). Melatonin use was not associated with an increased 5-year risk of HF (0.56% vs. 0.55%, sRR 1.01 [0.93-1.10]), ACS (0.55% vs. 0.56%, sRR 0.98 [0.89-1.06]), or IS (1.00% vs. 0.99%, sRR 1.01 [0.94-1.07]), Figure 1. However, melatonin users had a significantly higher 5-year risk of arrythmia compared with controls (0.80% vs. 0.71%, sRR 1.14 [1.06-1.22]). Subgroup analyses showed similar risks across subgroups, except for a slightly higher risk of arrythmia in those aged >45 vs. <45 years, Figure 2. Melatonin users had lower or similar 5-year risks of HF (sRR 0.50), ACS (sRR 0.88), arrythmia (sRR 0.66), IS (sRR 1.03) compared to Z-drug users. In the 1-year landmark analysis, a cDDD of >180 vs. 0 was not associated with HF (sRR 1.08), ACS (sRR 0.97), arrytmia (sRR 1.03), or IS (sRR 1.07).
Conclusion
Melatonin was not associated with increased long-term risks of ACS, HF, and IS. However, melatonin may be associated with slightly higher long-term risks of arrythmia. Further research is needed to elucidate whether this signal is a genuine safety concern or is driven by residual confounding.Figure 1Figure 2