Heart rate (HR) variability is impaired in chronic coronary artery disease (CAD), but the mechanism is not fully resolved. This study was aimed at assessing whether HR variability is influenced by the risk factors of CAD. Of a random sample of 120 subjects born in 1954, 88 (41 men and 47 women) could be included in the analyses. No subject had clinical heart disease. The subjects' physical activity, alcohol consumption and smoking were quantified by 2-month diary follow-up. Serum lipids and insulin were measured. The tests of HR variability included power spectral analysis and calculation of the root-mean-square difference of RR intervals at rest under controlled respiration. HR variability indexes were asymmetrically distributed and strongly HR-dependent, and therefore, all statistical tests were performed on log-transformed data adjusted to the population mean HR. Multiple regression analyses showed independent inverse relations between the root-mean-square RR difference and low-density lipoprotein (LDL) cholesterol (beta = -0.22; p = 0.008), and between the total RR-interval power and LDL cholesterol (beta = -0.25; p = 0.007), as well as smoking (beta = -0.19; p = 0.035). In women, alcohol use influenced the RR-interval root-mean-square difference (beta = 0.31; p = 0.015), total power (beta = 0.33; p = 0.017) and high-frequency power (beta = 0.26; p = 0.056). It is concluded that short-term HR variability is related inversely to LDL cholesterol and smoking in the population, and directly to alcohol use in women.(ABSTRACT TRUNCATED AT 250 WORDS)