Pinna · Clinical science (London, England : 1979) 2007 · test-retest reliability cohort study · n=39

Heart rate variability measures: a fresh look at reliability.

Cited 264 times in the scientific literature.

Level 3 - non-randomized controlled study

Prospective test-retest reliability study in a single cohort without clinical outcome endpoints

PubMed 17381425 · doi:10.1042/CS20070055 · record verified 2026-08-29

What was done

Researchers evaluated the absolute and relative test-retest reliability of standard short-term heart rate variability (HRV) metrics in 39 healthy adults (aged 26–56 years; 18 men, 21 women). Five-minute supine electrocardiograms were recorded during spontaneous breathing and paced breathing (15 breaths/min), and repeated the following day under identical conditions. Standard HRV metrics calculated were SDNN, RMSSD, LF and HF power (absolute and normalized), and the LF/HF ratio. Reliability was quantified using 95% limits of random variation and intraclass correlation coefficients (ICC), alongside sample size estimations to detect a mean difference of at least 30% of the between-subject standard deviation.

What was found

There were no significant group-level mean differences between the two recording days. However, individual-level absolute day-to-day variation was wide: the second measurement ranged from 0.5 to 1.9 times the first measurement for SDNN (best case) and from 0.3 to 3.5 times for LF/HF (worst case). Relative reliability remained high, with most parameters demonstrating an ICC > 0.8 (overall range: 0.65–0.88). Required sample sizes to detect moderate differences ranged from 24 to 98 participants. Paced breathing systematically improved reliability metrics.

Why it matters

While short-term HRV reliably ranks individuals relative to one another (high ICC), large day-to-day random swings in single individuals mean that individual monitoring requires substantial caution. Controlling respiratory rate with paced breathing reduces this noise and improves measurement precision.

Limits

The sample was relatively small (n = 39) and restricted to healthy adults across a limited age range, limiting generalizability to clinical populations with cardiovascular or autonomic disease. The retest interval was limited to 24 hours, so longer-term reliability was not assessed.

Cited by