Saturday, December 18, 2021

A tale of 2 HRVs - afib vs noise

In the last post I presented some facts and concerns about a potentially fatal arrhythmia, atrial fibrillation.  This past week I was asked to look over some HRV records that contained high levels of artifact to see if any data could be salvaged.  They each represent perfect examples of both the importance of proper "noise"/artifact diagnosis as well as the need for concurrent ECG recording (at least through the "noise" or artifact areas).  These examples come from a group of young, fit triathletes, all with very good VO2 max tests and no medical history.

Case 1:

Kubios plot of RRs recorded with Polar H10, during exercise at 90% VT1 power cycling:

  • We see that the artifact % is above 70% and the HR/RR are literally all over the map.
  • Is this noise, poor belt contact or some dangerous arrhythmia?

Here is the Movesense ECG (512 Hz) recorded concurrently:

  • It's perfect - no artifacts, arrhythmia and not even a single APC.
  • Obviously the H10 recording was at fault with some technical glitch or poor skin contact.


Case B (40 yo with VO2 max 51 ml/kg/min):

Kubios plot of RRs recorded with Polar H10 on two separate test recordings, same intensity as above:

  • Artifacts above 50% in both!
  • Same question - what is the underlying cause?

Movesense ECG:

  • No visible P waves
  • Irregular QRS, chaotic pattern, some pauses.

This is a 45 second ECG strip ran at 25mm/sec - a standard speed for electrocardiography:

I sent this to a friend and colleague who is an expert in ECG interpretation for review.  He felt this was most consistent with atrial fibrillation based on the chaotic, random pattern of QRS complexes and lack of noticeable P waves. Since it's only a single lead, the differential diagnosis still includes short runs of multiple APCs, or irregular SVT.  

The bottom line being the first case representing a noisy recording with no medical pathology and the second corresponding to a pathologic condition with major health consequences.  This is not as unusual as one might think.  In our first DFA a1 validation study, we had to exclude one participant because of atrial trigeminy (APC, normal, normal and repeats with APC...).  

Take home lesson - if you see substantial artifacts (with bluetooth), this may or may not be "just a poor connection" or motion noise.  

The Movesense ECG is a fantastic dedicated device to detect this type of problem, but let's not forget that using the Polar H10, Fatmaxxer will take an ECG snip of any artifacts detected and this can be viewed in Excel. Kubios or other software.