Why one night of HRV does not tell the whole story

Last night's overnight HRV read 70 ms, the second-lowest reading in a 14-night window whose low point was 68 ms and whose high was 94 ms. The reading capped a week that averaged 78.9 ms, down 9.2% from a prior week average of 86.9 ms. The same night's sleep score was 58, 4.9% under its own 7-day average of 61. Total sleep was 4.7 hours, with 0.5 hours of deep sleep and 0.5 hours of REM. Body battery at wake was 54, and resting heart rate was 52 bpm, 4% above its 14-night mean of 50 bpm.

What the plan called for

The training plan's fourth week called for threshold repeats at zone 4, at an intensity of 0.95. The day's status in the plan log was matched. The logged session ran 52 minutes with an average heart rate of 134 bpm, and its training effect was labelled speed. There was one session that day, and it was not logged as hard.

Why one number misleads

A review of HRV monitoring in elite endurance athletes found that both rises and falls in HRV have been linked to negative adaptation, which is why a single nightly reading misleads.¹ The authors recommend appropriate averaging techniques and longitudinal monitoring, so an athlete's own pattern over time replaces any one number.¹

A 4-week running study put this into practice. Twenty-six healthy, moderately fit men were split into two groups: nine had their training set by daily HRV measurements, and eight followed a predefined plan.² The HRV-guided group raised VO2peak from 56 to 60 ml/kg/min (P = 0.002); the predefined group's VO2peak did not change significantly, from 54 to 55 ml/kg/min.² The change in VO2peak did not differ significantly between the two groups. The HRV-guided group was prescribed easier training or rest when HRV fell below a reference value based on the 10-day mean and standard deviation, or showed a decreasing trend for two days.²

A later 8-week study of 40 recreational runners compared the same two approaches.³ The HRV-guided group trained with fewer moderate and high-intensity sessions than the fixed-plan group, 13.2 against 17.7 on average (P = 0.021), and improved 3000 m performance by 2.1% (P = 0.004); the fixed-plan group's 1.1% gain carried a P value of 0.118.³ The study's authors describe the difference between the two groups as small.³

A meta-analysis across randomized trials in trained endurance athletes found that both HRV-guided and control training improved VO2max (p < 0.0001), with a significantly higher effect for HRV-guided training, 0.187 above control.⁴ The overall effect of HRV-guided training on VO2max was itself small, 0.402, and the authors say it depends on the athlete's level and sex.⁴

A broader meta-analysis of 8 studies and 199 participants found HRV-guided training better improved vagal-related HRV measures than predefined training (SMD 0.50, 95% CI 0.09 to 0.91), while differences in resting heart rate, maximal aerobic capacity and endurance performance were not significant.⁵ Its authors write that any advantage for fitness and performance is, on current data, by a small margin.⁵

What the numbers say about last night

The brief's own two-week statistics put last night's HRV at a z-score of -1.56, using a 14-night mean of 82.9 ms and a standard deviation of 8.3 ms: more than one standard deviation under the average, but short of the window's lowest reading. Sleep score sat at a z of -0.66 against a 14-night mean of 63.7, and resting heart rate at a z of 1.06 against a mean of 50 bpm. None of the three crossed two standard deviations.

Sleep score tells a similar week-over-week story: a last-7-night average of 61.7 against a prior-7-night average of 65.7, a drop of 6.1%. Resting heart rate moved the other way, up 4.5% over the same comparison, from a prior-7-night average of 48.9 bpm to a last-7-night average of 51.1 bpm.

• Compare the last 7 nights against the 7 nights before them.
• Read one low night as a single data point inside a longer trend.
• Keep a session the plan calls for unless the trend holds across several nights.

Does a low HRV reading mean training should change today?

The 2007 HRV-guided training study eased sessions when HRV fell below an athlete's own 10-day reference range, or trended down for two days running.² That decision came from a reference range built over days. Last night's HRV was low against the 14-night average, and the wider week ran 9.2% under the week before it, which is the fuller picture.

How many nights of data does it take to see a real trend?

The studies behind HRV-guided training used windows such as a 10-day mean and standard deviation, or two straight days of decline, to decide on easing training.² A 14-night window, like this one, is enough to tell whether a dip is part of a week-long pattern or a single off night.

Reading your own trend

None of this says a low HRV night should be ignored. It says the useful signal sits in numbers built from several nights: the 7-day average against the 7 days before it, the standard deviation around a two-week mean, and whether a dip holds for more than a night or two.