Reproducibility of post-exercise heart rate recovery indices: A systematic review

Heart rate recovery (HRR) has been widely used to evaluate the integrity of the autonomic nervous system with a slower HRR being associated with greater cardiovascular risk. Different HRR indices have been proposed. Some evaluate HR changes from the end of exercise to a specific recovery moment (e.g. 60s - HRR60s; 120s - HRR120s; 300s - HRR300s) and others calculate time-constant decays of HR for different recovery intervals (e.g. first 30s - T30; the entire period - HRRt). Several studies have examined the reproducibility of these commonly-used HRR indices, but reported discordant findings. Thus, this systematic review was designed to synthesize the reproducibility of HRR. We included studies that evaluated short-term (<1 year) reproducibility of HRR after dynamic exercise by employing typical measures of reliability (intraclass correlation coefficient, ICC) and agreement (coefficient of variation, CV). The electronic database PubMed/Medline was searched for relevant studies published up to July 2018. From the initial 120 records identified, 15 studies were retained for the qualitative synthesis of 24 experimental conditions. During most experimental conditions, high ICC and desirable CV were reported for HRR60s (62.5 and 76.2%, respectively), HRR120s (55.6 and 71.4%) and HRR300s (50.0 and 100.0%). While, it were reported during the minority of conditions for HRRt (37.5 and 42.9%) and in none condition for T30 (0.0 and 0.0%). In conclusion, HRR60s, HRR120s and HRR300s exhibited good reproducibility for evaluating HRR in predominantly healthy males within research and clinical settings. In contrast, caution should be taken when employing other HRR indices (T30, HRRt) due to their poorer reproducibility.