The mean change of RHR was − 1.1 beats/min between two consecutive visits for a routine health examination (mean 2.9 ± 1.7 years between visits). As expected from the literature, women presented higher RHR than men (73.0 vs. 69.8 beats/min, p < 0.001), lower BMI and blood pressure, lower MET and improved lipid profile (all p < 0.001), but the ∆RHR was similar between sexes (p = 0.925).

Population characteristics presented in Table 1 and the distribution of ∆RHR is shown in Fig. 1 for men and women, with x-axis reference lines dividing the cohort to quartiles of ∆RHR, so that patients in the first two quartiles (left side of the distribution) had lower RHR at the follow-up visit compared to the baseline visit (their RHR decreased with age) while the RHR of those on the right side of the distribution was higher at the baseline measurement than on the follow-up measurement (increased over time).

RHR was very weakly correlated with age and BMI (r = − 0.06, r = 0.07 p < 0.001) but moderately correlated with MET (r = − 0.39, p < 0.001). Controlling for MET revealed stronger negative correlation between RHR and age (r = − 0.205, p < 0.001). These results remain significant when we split the analysis by sex.

Next we adjusted RHR for sex, BMI, and MET and plotted the residuals against age (r = − 0.221, p < 0.001), meaning that even after controlling for these 3 confounders, RHR decreased with age (Fig. 2).

In this cohort study of healthy, employed adults without known cardiovascular disease, we found that RHR decreased with age, even after adjustment for sex, BMI and MET. We found that RHR values tend to decrease with aging (mean change in RHR was − 1.1 beats/min per 2.9 years of follow-up), so that RHR increases are the exception, and chose to focus on the correlations for this change, emphasizing the importance of recognizing increased annual heart rate as another risk factor for cardiovascular disease and mortality. We showed that age, female sex, baseline RHR, HDL-c and MET have a beneficial effect on heart rate change (decreased RHR), while increasing body weight (expressed as ∆BMI) and presence of dyslipidemia have an adverse effect (increased RHR).

Elevated RHR has been established as a risk factor for cardiovascular morbidity and mortality. Some studies have shown RHR to be predictive of all-death mortality and some link RHR to malignancies [7‐9].

Our study found that RHR decreased with age. The effect of age on RHR is less well-established than that of other non-modifiable determinants [23‐25]. In some studies there appears to be a decrease in RHR with age [12, 26‐29]. Some show no change in RHR or a decrease followed by a plateau [12, 29‐32]. Others found a decrease in RHR in women but not in men [8, 30]. Palatini et al. found an increase in RHR with age, in patients with isolated systolic hypertension [33].

We found annual increased heart rate to be correlated with a decrease in MET. The metabolic equivalent of the task (MET) is commonly used to express the oxygen requirement of the work rate during a stress test, and demonstrates levels of fitness. Decreased MET, therefore, implies less efficient heart function and worse cardiovascular fitness.

Generally, cardiac output and heart rate decrease with age [34, 35].), therefore, individuals with an increasing trend of RHR should be encouraged to choose and maintain a healthy lifestyle in order to decrease cardiovascular risks. Nevertheless, the clinical significance of individual ∆RHR remains to be examined in larger-scale cohorts.

Furthermore, our findings highlight the importance of fitness and heart rate tracking. The development in personal digital devices has made heart rate information as important and easily access as ever, thus, provides a golden opportunity to identify use delta heart rate as an early marker of disease prevention. However, currently there are no general recommendations to guide the general public in this area [36].

Heart rate is controlled by the autonomic nerves, with sympathetic stimulation increasing the rate and parasympathetic stimulation decreasing it. Athletes have lower RHR compared with the general population and training results in reduced RHR. This is explained by changes the sympathovagal balance of the sinus node. In response to regular exercise changes occur in the cardiovascular system, such as increased contraction due to cardiac muscle fibers hypertrophy and increased muscle mass of the ventricles. Oxygen and nutrient delivery to muscles is improved by enhanced capillary capacity for blood flow, which causes a decrease in total peripheral resistance [37].

The importance of fitness to health is well-known and studied, and there is an inverse relationship between fitness and mortality. Kokkinos et al. examined MET in 18,102 men who were followed for a median of 10.8 years, and found that for each additional in exercise capacity of 1-MET, mortality risk was reduced by 12% [38].

We also found annual increased heart rate to be correlated with increase in BMI. The HARVEST study found ∆RHR to be an independent predictor of the development of weight gain in young persons screened for stage 1 hypertension, and several studies link increased RHR and metabolic syndrome.

High BMI is a well-known cardiovascular risk factor, even in metabolically-healthy overweight people [39]. Interestingly, the increase in BMI seen in our study population remained in the normal range of BMI and our cohort was not over-weight or obese. Therefore, increase in RHR was affected by increase in BMI even in individuals with normal BMI.

This study analyzed the correlations of annual ∆RHR; it does, however, have limitations. While initial study population included 18,083 subjects after applying exclusion criteria (see “Methods”), at the second time point we were able to follow-up on only 7735 (42%) subjects, who actually arrived at our center for annual routine checkups. We systematically checked for non-response bias and found that non-participants did not differ from participants on any of the socio-demographic or bio-medical variables. Furthermore, a high rate of participation was observed (91.6% of those who arrived at the center and were asked to participate). Another limitation is that our cohort consisted of participants in a health screening program and is not a population-based sample. However, the mean BMI in our study is very similar to that published by the National Health Survey from Israel [40]. Second, mean follow-up was 3 years, which is a relatively short time period for the investigation of RHR trajectories. Nevertheless, we were able to show a significant link between the biomarker studied and RHR trends in this cohort of apparently healthy individuals. We expect that the observed trend will increase with time unless modification in life style will be done. Another possible limitation is that the use of baseline RHR measurements introduced potential bias with respect to regression to the mean. Suggesting that individuals with extreme RHR on its basal measurement but closer to the mean on its second measurement [41]. The very large spread in the change in RHR (Fig. 1) are most likely not due to biological factors, but other factors such as measurement error, change in measuring circumstances, regression towards the mean, and others. Last but not list, it was previously shown that among employees, white collar workers (professional, managerial, clerical) had slightly higher resting heart rates than blue collar workers. This suggests a possible effect of greater physical fitness among blue collar workers [42]. However, in our study we did not collect data regarding type and amount of physical activity during work hours.

Our results confirm that RHR decreases with age and strengthens the need for identifying patients with annual RHR increase as a population at risk. Determinants of increased annual RHR over time and its consequences include initial lower exercise capacity, dyslipidemia, increase of BMI or decrease in MET.

Our findings may be useful in identifying easily, and without any additional cost or time, asymptomatic individuals at risk, who could benefit from primary prevention (lifestyle changes or medication) in order to reduce cardiovascular risk.