The aim of this study was to assess the effects of 5 mg melatonin before sleep in patients with coronary artery disease (CAD) and with an abnormal circadian pattern of blood pressure (BP) on changes in circadian BP profile and heart rate variability (HRV). Sixty patients with CAD, nondippers aged 48-80 years (male 75%), were included. In addition to previous treatment, they were randomly allocated to melatonin or placebo. After 90 days, a second 24-h BP monitoring was carried out. Each patient had two sessions (before randomization and at the end of study) of 24-h ECG monitoring to assess the changes in HRV. Inclusion of melatonin led to BP pattern normalization in 35% of patients in the melatonin group and in 15% of controls (P=0.609). This effect was reached not only by a decrease in nighttime BP, but also by an increase in daytime BP (significant in the melatonin group). A nonoptimal effect for BP profile was observed in 12.5% of patients: extreme- or reverse dippers. In patients with conversion from nondippers to dippers (responders), an increase in standard deviation of normal-to-normal intervals between initial and final HRV analyses was observed. Nonresponders represented an increase in the mean circadian heart rate. To avoid nonoptimal effects, the inclusion of melatonin in pharmacotherapy of patients with CAD should be based on monitoring of circadian BP profile, before and during treatment. As melatonin caused not only a nocturnal decrease in BP but also a daytime increase, it should not be recommended in patients with \'high normal\' values of BP because of the danger of induction of arterial hypertension.