Saturday, June 8, 2019

Firstbeat VO2 estimates - exercise in the heat

Does the derived VO2 remain stable in diverse conditions of temperature?

I recently visited my son in cool clear San Francisco, leaving the warm Florida weather behind.  Before leaving, my VO2 max was a calculated 50 by 4 minute maximum power and the Garmin derived VO2 max was set at 49 ml/kg/min.  On the day of travel I did 30 minutes on an indoor trainer at home and the VO2 rose to 50.  Two days later I did 60 minutes on trainer outside (72F) with 1 MLSS interval and the VO2 max went to 51.  Two days after that was another trainer ride (but at 52F) with 2 MLSS intervals and VO2 max was displayed as 52.   When I returned back to Florida the heat and humidity kicked in and my heart rate skyrocketed.  The VO2 numbers were also dropping down somewhat.  This lead me to look both at the heart rate/power trends as well to review the literature in regards to VO2, heat, heart rate and stroke volume. 

The last post on VO2 estimation utilizing both maximal and submaximal tests introduced several concepts.  The VO2 (oxygen usage in ml/kg/min) is related to both the cardiac output (stroke vol x heart rate) and the oxygen drop across the active muscle (arterial- venous O2 difference).  Therefore, at a maximal VO2, the cardiac output and the muscular oxygen extraction should both be at their peak potential.  Predictive equations based on maximal running speed or bike power for 4-5 minutes equate oxygen usage to the amount of power used.  This makes sense since the faster one goes, more power is needed and more fuel will be burned.  The validity is also dependent on the particular exercise efficiency factor being similar across populations. In sports such as cross country skiing where optimal technique is essential, it may be more difficult to derive a VO2 value based on performance over a maximal 5 minute course.

Predicting the VO2 max based on submaximal efforts such as Firstbeat or Astrand algorithms make several different assumptions. The main assumption of a submaximal calculation of VO2 max is that there is a linear relation of VO2 to heart rate between certain loads.  This has been shown to be true up to the MLSS power.  The basis of the Astrand single stage test was founded by his research that 50% of the VO2 max was reached at a heart rate of 128 bpm for men (138 for women).  He then extrapolated to find the max or 100% VO2 value using this measure.  Given that there is a linear relation of heart rate to VO2 (in the low to high-moderate power range), factors that can effect heart rate like age,  temperature, humidity, hydration, nervousness, caffeine, will all modify the predicted VO2.  

In an effort to factor out heart rate related variables, Ekblom-Bak suggested looking at the HR at 2 different power levels based on the evidence that subjects with higher VO2 fitness had a blunted HR to Power slope:

Depending on both the absolute and relative increase in heart rate as power is increased, a better idea of VO2 peak can be obtained.  

The above technique is presumably also being used by Firstbeat in their VO2 estimation.  Under most circumstances it will usually be sound from a physiologic standpoint.  

Let's look at the sequence of the training sessions.  Even though an outdoor Florida ride may be at a similar temperature as an indoor, the humidity outside makes the "feels like" temperature about 10 degrees higher.

Pre trip Trainer:

Outdoor trainer in SF, warm outside:
  • An extra few watts leads to a 3 BPM increase in heart rate.
  • Rectus Femoris SMO2 slight down-slope signifying MLSS exceeded.

Outdoor trainer in SF, cool outside about 52F:

  • Similar heart rate and power from day prior, but higher VO2 value.
  • The watch temp was 63 vs 80F.
  • RF SMO2 pattern the same.

Back to Florida:
Outside temp 78F with high humidity:
  • Much higher heart rate despite lower power.
  • Stable RF SMO2 (at MLSS).
  • VO2 still same at 50.

Inside trainer two days later:

  • Some heart rate elevation at similar power.
  • Minimal SMO2 drift down at about MLSS plus.

Here is a summary:

                      Temp (F)         Watts     Last min HR    VO2 max                           Day
Pretrip              80                  255           145                 50                                    1
SF #1               72                  257           148                 50                                    3
SR cool           52                  257           147                 52                                    5
FL hot             90                   253           156                 50                                   9
FL indoor         78                   253           150                 50                                  14
FL hot              86                   252           156                 50  (graph not shown)   16

  • The overall stability of the Firstbeat estimation is pretty amazing considering the variation of heart rate at the same power as I've shown. 
  • The bottom line here is despite very different riding conditions, the calculated VO2 was remarkably stable.  This is a huge accomplishment and my compliments to Firstbeat for the hard work that went into this.  

Some question remains though:
  • Does relative heart rate elevation from exercise in the heat effect the true VO2 max (as the actual gas exchange gold standard)?
  • Will the VO2 max derivation from power/heart rate kinetics still be valid, since it is based on certain assumptions such as "predictable" stroke volume change?

To start with some "normal" physiology, the stroke volume does not always stay the same as heart rate rises.  This is important to factor into the VO2 estimate since the VO2 relation to heart rate assumes the stroke volume not to change with high loads.  From the paper by Rowland, several studies show that endurance athletes have a progressive rise in SV with heart rate/load:

However, not all studies show this and the conclusion of his review was:
Whether or not endurance athletes demonstrate
a progressive rise in stroke volume during a
progressive exercise test, and the mechanisms
surrounding their generation of superior cardiac
output differ from those of untrained individuals
cannot be satisfactorily answered based on the
current body of research data. From this review,
it is apparent that equivalent amounts of evidence
can be mustered to support both sides of the
argument. The available information suggests the
possibility that testing methodology and protocol
might be responsible for the marked variation
in stroke volume observed in these reports. If a
non-plateau in stroke volume is typical of elite level
endurance athletes, enhanced diastolic
filling would appear to be the most likely mechanism
with a reliance on a Frank-Starling
mechanism to augment stroke volume as work
intensity rises
I have corresponded with the Firstbeat representative and he did indicate that a stroke volume correction is present in their formula.  This is certainly good news, but there will probably be individual variation in the amount that the SV alters.

Interaction of heat and heart rate
This where the issue of cardiac output, VO2 and stroke volume can become troublesome.  A very well done study was just released -
"Low Stroke Volume during Exercise with Hot Skin Is Due to Elevated Heart Rate"
The purpose of the study was to investigate the mechanism of the progressive decline in stroke volume seen during hyperthermic exercise.  The prevailing theory for this is that high cutaneous blood flow and cutaneous venous pooling (redistribution of blood volume for skin cooling) leads to decreasing central venous return and a lowering of cardiac filling pressure.  The reduced stroke volume (from the low filling pressure) in turn is compensated for by a higher heart rate.  The alternate possibility is that heat stress itself causes excessive cardiac rate drift upwards (as a primary process), leading to a shortening of the time that a chamber can fill with blood.  
To test which effect is actually happening the authors gave low dose beta blockers to prevent the heart rate rise in hot conditions (so same net heart rate in control vs hot subjects).  The results are fascinating:

  • You can see that the cardiac output (top graph) is lowest in the hot group (despite the highest heart rates).
  • The best stroke volume is in the beta blocker group but the overall cardiac output is the same as control since heart rate is a bit lower. 

The cutaneous flow and volume were increased as well:

There was no relation between venous forearm volume with stroke volume: 
The conclusion of the paper was:
Increased Tsk and Tes raised CO and CBF through a disproportionate increase in HR (almost to maximal) which in turn directly lowered SV. Increased Tsk and Tes also lower TPR and MAP, assumedly due to increased CBF. No one variable appears to be primarily regulated, yet there appears to be systematic compensation. For example, the increase in CBF is met by increased CO which is accomplished by increased HR. The increased CO also minimized the drop in TPR and MAP. The present findings also point out the robust nature by which CO can be achieved; through a disproportionately high HR in HOT or a disproportionately high SV in HOT-βB.

In conclusion, the relatively low SV observed during moderately intense exercise when Tsk is hot (38°C) can be explained largely by tachycardia and shortened diastolic filling time. Lowering HR with βB restored SV to the same level as when skin temperature was 33°C and occurred independent of measurable effects on CBF, CVC, or cutaneous venous compliance
Stroke volume vs heart rate interaction expanded:
The comment above - "The present findings also point out the robust nature by which CO can be achieved; through a disproportionately high HR in HOT or a disproportionately high SV in HOT-βB." is observable in a more general fashion as well.  Namely that the product of stroke volume x heart rate seems rather predetermined for a given individual at a given point in time.  In other words, total cardiac output can be manipulated by an atypical rise in HR or SV but not both.  How did I come to this conclusion?  Another great study done some years back looked at MLSS, VO2 max in normal subjects on and off beta blocker therapy.  The initial expectation was that using the beta blocker would cause a decrease in submax and max heart rate, leading to a reduction in cardiac output and the above performance metrics.  In fact, I was always taught that B blockers should be avoided in patients who desire maximal aerobic performance.
Subjects randomly received oral bisoprolol (5 mg/day) or placebo
according to a randomized, double-blind crossover protocol. Each
treatment period lasted at least 2 weeks, with a washout phase of
1–3 weeks between the two treatment periods. Tests were performed
in the 2nd week of each period
Baseline characteristics of test subjects at rest:
  • The lower resting HR is most obvious, other parameters are the same.


  • The heart rate at MLSS was 23 bpm down in the beta blocker group despite similar power, ventilation and lactate.  
  • The oxygen pulse (an index of stroke volume) was higher in the beta blocker group which presumably lead to a similar net cardiac output in both groups.

The VO2 max:
  • We see a similar situation as the MLSS.
  • Heart rate is much lower in the beta blocker group by 19 bpm but power output, ventilation, and VO2 max was the same.
  • Oxygen pulse (stroke vol) was higher with beta blockers.

Therefore an individuals net cardiac output potential is relatively fixed (at a given point in their training) and is achieved by the interplay of heart rate elevation and stroke volume compensation.  

Does maximal heart rate improve with training and fitness?
Is the max heart rate a trainable trait like stroke volume, mitochondrial number or strength?
It does not appear to be.  The estimate of 220 - age also appears to be incorrect.
According to the HUNT study:
The interaction between age and tertiles of VO2peak was not significant among men (P = 0.19) or women (P = 0.19).
Age was the best predictor with the resultant equation:
HRmax was univariately explained by the
formula 211 - 0.64·age (SEE, 10.8), and we found no evidence
of interaction with gender, physical activity, VO2max
level, or BMI groups
However, even though the fit subjects will not have a higher max heart rate, they do have the blunted heart rate rise with increasing load so can achieve much higher maximal power outputs.  Long term endurance training does enhance stroke volume so even if the max heart rate is no different, the overall cardiac output is certainly better.
I have extrapolated the Ekblom Bak drawing to show this:

The reason I present these studies is that the Firstbeat formula is heavily based on heart rate monitoring but also on an assumption of stroke volume. The interaction is obviously complex so a one size fits all approach may lead to errors.  

In addition, it appears that there area complex group of compensatory changes that occur with heat and dehydration to reduce time to exhaustion as well as VO2 max.  Here is a figure from an excellent review:

Final thoughts on getting an accurate Firstbeat VO2 max calculation:
  • Measuring equipment must be of high quality and relatively free from artifact.  The heart rate monitor should not be optical.  Drop outs from your power meter or poor GPS accuracy for runners can invalidate the results.
  • From trial and error as well as personal communication from Firstbeat, I believe the following types of intervals are optimal to VO2 formula accuracy - 5 to 10 minute continuous intervals at the MLSS, maximum constant power 3 to 5 minute intervals.  Doing low level constant pacing may also be helpful to provide further data points to the heart rate power relationship.
  • The formula seems to handle heat induced heart rate elevation fairly well.  Although it may not compensate fully, the true VO2 max may really be down depending on the severity of the hyperthermia.

  • Estimation of VO2 max at less than maximal intensity is based on the relative linear relationship of heart rate to VO2 through a range of power intensities.
  • This extrapolation relies on the assumption of a similar stroke volume as intensity increases.  A rise in stroke volume can be seen in endurance athletes to a variable degree, but not every study agrees with this change.  There may be intra individual variation in the stroke volume, work rate relationship.
  • Firstbeat equations, take the athlete's stroke volume change into account in their calculation (personal communication).
  • Exercise in the heat leads to heart rate elevation at similar power intensities.  Although there is an increase in cutaneous blood volume, it's not primarily related to stroke volume reduction.
  • The heart rate elevation that occurs while exercising in the heat will lead to shorter times of cardiac filling and a reduction in stroke volume.  Cardiac output is preserved as a consequence of the higher heart rate.
  • Beta blocker therapy in a normal subject will not result in a change in VO2 max or MLSS.  Although heart rate is lower in this instance, the longer cardiac filling time leads to greater stroke volume and preserved net cardiac output.
  • Beta blocker therapy will prevent the heat induced rise in heart rate during exertion, prevent the fall in stroke volume and normalize cardiac output to that of a normal thermic state.
  • The maximal heart rate of an individual is not related to VO2 max and is not a trainable feature.  
  • The Firstbeat VO2 max estimation relies upon accurate heart rate and power measurements.  Failure to use "gold standard" heart rate devices will result in erroneous results.
  • Despite the complexities of temperature related heart rate interaction and stroke volume , the Firstbeat calculation was remarkably stable for me.  

 VO2 max related posts
    HRV related posts


    No comments:

    Post a Comment