Monday, April 23, 2018

Issues in using Muscle O2 for lactate testing

 An update on using NIRS for MLSS determination using long intervals
Additional data on using both NIRS and blood lactate for MLSS

The transition of steady state levels of lactic acid to significantly higher amounts is known as the lactate threshold.  The power output associated with the chemically defined level of decompensation has bearing in training load selection as well as a benchmark test parameter.  The search for non invasive surrogate markers has lead to the use of NIRS/muscle O2 as an alternative.  In fact both Moxy, BSX and now the Humon sensor have study data supporting this usage and accuracy.
However there has not been universal acceptance of this correlation, especially in view of what muscles are monitored, depth of O2 sensor measurement and site on the muscle.  Recently an interesting paper was published taking another look at this situation, namely can the Moxy sensor tracing truly estimate lactate or the ventilatory threshold.  A word about the ventilatory threshold.  As lactate accumulates with intense exercise, lactic acid dissociates to lactate and H+.  The H+ is buffered by bicarb and through several steps eventually leads to production of CO2 and water.  This rise in CO2 (above the usual CO2 from metabolic processes) leads to higher minute ventilation that can be measured and compared to the overall O2 consumption(VO2).  This ratio's breakpoint is the ventilatory threshold.
The study measured the lactate, ventilatory thresholds as well as the muscle O2 transition using a Moxy sensor on the VL during a graded ramp protocol.


As noted the thresholds were quite different on each test type.
Looking at the individual data points showed even more interesting insights:

The potential difference (erroneous threshold for that person) between the Moxy and blood lactate was quite high, +- near 50 watts.  Some participants were very close in NIRS vs blood lactate power, but how to identify those beforehand will be problematic.
The correlation was better with the ventilatory vs lactate figures.

The bottom line is that a Moxy derived lactate threshold can be up to 50 watts off that of a blood measurement.

Note, I'm not picking on the Moxy, I'm sure the Humon or BSX units would have yielded similar resultsIt's the process not the sensor.

The authors make several logical reasons for the wide variation in TSI/muscle O2 thresholds in relation to blood lactate.
  • The lactate measurement is systemic and presumably is affected by both locomotor, respiratory and "other" oxygen consuming tissues.  The muscle O2 on the other hand is measured at a relatively small site, in one muscle and at a shallow depth.
  • There may be different behavior in proximal vs distal muscle sites.
  • Muscle O2 measurement inaccuracy from variable thickness in adipose tissue.

For interest, it was felt that the ventilatory threshold occurs late, after lactate has already risen.  This is due to the time needed for the central chemo receptors to react to the multiple chemical steps needed and blood brain barrier penetration.  Therefore not very helpful for real time use since it occurs after the fact.

Finally I will end with the articles conclusion:
Therefore, it would appear that neither TSIT nor VT are suitable non-invasive determinant of the aerobic-anaerobic transition when both are simultaneously compared to the invasive measure of TLac

Next posts will get back to costal O2 usage, use in a prototypical endurance athlete(not me) and SNP genetic markers

 Lactate related posts


  1. Hi, thanks for posting this study - it is really useful but I think the authors made some errors!

    It looks to me like they have confused a number of thresholds in their study and then discussion. It looks like TSI is in fact correlated with aerobic threshold or VT1 ie the first rise of lactate above baseline [225-250w step in a)], this is quite good to know. OBLA or anaerobic threshold or VT2 should be around 4mmol which indeed is around 50w higher.

    Their lactate threshold calculation which confuses both aerobic and anaerobic thresholds looks to be the issue. Thanks for posting the paper.

    Kind regards,


  2. Yes, they use the VT rather loosely at times which is confusing.
    I think the most useful piece of data is Fig 3, the Bland-Altman plot showing high variation per subject. Their point may be that the error can be high, in whatever threshold you want to ascribe to.
    The protocol was also interesting, 25 watts every 3 min on the ramp which is almost a steady state escalation. I am currently working on a post looking at the O2, HHb tracings just below and above MLSS and (at least my data) it is very intriguing for displaying breakpoints.