Monday, 23 May 2016

Illness and HRV

Stress is a vital part of athletic training. But too much means you don't have time to recover before the next session and spiral yourself into exhaustion, whilst not enough means you're never working hard enough to improve. Many are touting Heart Rate Variability (HRV) as a potential solution.

Most decent athletes try to monitor what they're doing to themselves. People use TRIMP (heart-rate-based, related to Strava's Suffer Score), TSS (power-based), RPE (effort-based) or a variety of other measures to quantify training load. With some idea of how much we can tolerate, we can try to keep increases below a level that might produce exhaustion.
However, we all get ill - especially when we're training too hard. A cold or other infection is another 'stress' to the body which isn't accounted for by any of the 'bottom-up' methods of quantifying stress. Is there another way to measure it?


One way is to use Resting Heart Rate (RHR). Measuring RHR first thing in the morning can indicate illness as it rises when the body is fighting off infection. It's well established; we use it in hospital; but it's a bit of a blunt instrument. It takes quite a bit of infection to put it up, and even when it does rise it may not be by much (10% with a heart rate of 40 is only 1 beat in 15 seconds!).

I've been using Heart Rate Variability (HRV) in various forms for a while now and think it might offer something extra. And (luckily?) I've been unwell a couple of times over the last few months and have some data to show you.


I use ithlete, combined with a Polar H7 chest strap to measure HRV every morning. Other apps exist, some of which are free, but I quite like ithlete for its easy logging of a daily value and tracking of weekly averages. It uses the 'RMSSD' metric (for geeks it actually reports 20 x ln(RMSSD) ) which to me makes a lot more sense than mucking about with Fourier analysis and arbitrary frequency cutoffs. And has the advantage of producing a number that looks vaguely like a percentage. I take it standing each morning (if I lie down it goes over 100 which I gather reduces the sensitivity somewhat) and find it's normally around 80-90 if I'm well and rested. It's not by any means perfect - kids running around, dehydration etc can all alter it - but looking at the averages seems to show something.

My HRV with a chest infection
In mid-March I started with a nasty chest infection which went on for weeks. Initially I wasn't sure whether it was just a bad cold - but when I woke up with an HRV of 48 I was a little concerned! As it's a logarithmic scale a loss of 40 points means about a 7-fold loss in variability. It wasn't for 4 weeks that it was regularly back in the 80s again which nicely matched with how long I felt washed-out for.

What did I do during that time? For the initial very-low phase shown above - nothing. No training. To be honest I didn't feel like it. When it started creeping up into the 70s I started some basic easy (Z1/Z2) work - I find it's the intervals that usually cause HRV to crash the next day. It was painful spending a month like that, but I managed to start the season properly with an unexpectedly-good performance at the ETU Long Distance Duathlon so something evidently worked! (On the downside, the complete lack of speedwork was something I really felt in the race!)

I'll write more on HRV and ithlete in the future, but I thought the above was useful to show it in action.


No pretty graphs to show here, but it's worth knowing that everyone using a recent Garmin watch (I have the 920xt) also has HRV available to them. There's a ConnectIQ app if you want to measure it yourself in the same way as ithlete above, but Garmin use algorithms from FirstBeat for 'Training Effect' and the 'Recovery Adviser' as well as attempts to report VO2Max and Lactate Threshold. These are mostly based on heart rate and HRV.

I'm not a big fan of 'Training Effect' - though I'm disappointed if I finish a race and it's not near or at 5.0. It seems to value long, 'quite hard' sessions over shorter harder intervals despite what my legs say about them. However, the recovery adviser seems to get at something interesting in me. After a relatively easy session the adviser will recommend rest in the 10-20hr region. After a hard one it's usually 20-40hrs. Anything over 40hrs suggests I've gone really hard (or a race).

What I've discovered is that an early sign I'm overdoing it (or coming down with something) is the recovery adviser suggesting something that seems way out of keeping. A 1hr Z2 turbo that gets me a 25hr recommendation, for example. That's usually backed up by my HRV dropping the next morning.


HRV - well worth a look in this domain (illness). If you're using a recent Garmin watch you may have it already. ithlete is a simple way to get started with HRV - if you want to really explore it try something like CardioMood. I'll write more on my thoughts on HRV (particularly around ithlete) in the future.


  1. 48!?..Jings..was there no indication prior to that number that something might be going on as tha's one hell of a drop

  2. The day before I'd felt like I might be starting something in the afternoon (vague fuzzy head really), but thought it might just be a cold or a response to training hard in the morning. It was a couple of days before the real chest symptoms started.

  3. I thought your suppose to do a reading sitting up? Or does it matter? I know in the hospital i was lying down and i had to do 3 forceful breaths like you do for a lung test

  4. Most important thing is to be consistent from day to day. However, if you have a low resting heart rate then there's a theory that you can have 'parasympathetic saturation' which means that your HRV won't drop so much from stress like illness or overtraining. Taking the reading standing up adds just a little bit of stress (as your heart has to pump 'uphill') and avoids the saturation.
    I tried recording both lying and standing for a while; I certainly found for me that the standing seemed to be more responsive to changes so that's what I've stuck with.
    Resting heart rates are usually recorded lying as that's when it's lowest.

  5. Training effect is on the nervous system. The concept of a low score for long slow runs being you're nervous system should recover esp if heart rate was kept low. I found this true in marathon training where I built to 3 hour 20 mile runs on a Monday morning and had some of my best club sped sessions on a Tuesday night...

    1. I agree - I find a low score can be markedly improved by a long slow session (I usually use the bike rather than running) and a good night's sleep.