The prognostic implications of heart rate (HR) change over years have never been assessed. It was hypothesized that an increase in HR in apparently healthy persons observed over years could be associated with an increase in mortality risk and conversely. A total of 5,139 asymptomatic working men (aged 42 to 53 years) free of clinically detectable cardiovascular disease were recruited from 1967 to 1972 and had their HRs measured at rest in standardized conditions every year for 5 consecutive years. HR change was defined as the difference between HR at examination 5 and HR at inclusion, and subjects were divided into tertiles according to decrease >4 beats/min, unchanged (from -4 to +3 beats/min), and increase >3 beats/min. After >20 years of mortality surveillance, 1,219 deaths were observed. After adjustments were made for confounding factors, including baseline HR at rest, and compared with subjects with unchanged HRs, subjects with decreased HRs during the 5 years had a 14% decreased mortality risk (RR 0.86, 95% confidence interval 0.74 to 1.00, p=0.05), whereas subjects with increased HRs during the 5 years had a 19% increased mortality risk (RR 1.19, and 95% confidence interval 1.04 to 1.37, p<0.012). In conclusion, change in HR at rest over 5 years was an independent predictor of mortality in middle-aged men.