Gear & Technology, Health & Injuries, Podcast, Science and Physiology

Muscle oxygen saturation and iliac artery flow limitation with Jem Arnold | EP#358

 October 3, 2022

By  Bernardo Gonçalves


Jem Arnold - That Triathlon Show

Jem Arnold is a physiotherapist, coach and PhD student. Currently, his main research focus is investigating FLIA: blood flow limitation of the iliac artery, a condition that affects certain athletes even at the very elite level, sometimes with surgery as the end outcome. Jem is also very involved and knowledgeable in using NIRS (near-infrared spectroscopy) to measure muscle oxygen saturation (SmO2). We dive deep into how using NIRS and SmO2 can inform workout execution as well as training interventions and athlete profiling.

In this episode you'll learn about:

  • What is NIRS and what is muscle oxygen saturation?
  • The balance of oxygen delivery (to the capillaries surrounding muscle tissue - supply) and extraction (by the muscles from the capillaries - demand)
  • NIRS testing protocols, and how it relates (and does not relate) to oxygen uptake (VO2) and lactate measurments
  • Using NIRS to profile athletes
  • Using NIRS in high-intensity workouts to optimise interval execution
  • Blood flow limitation of the iliac artery (FLIA) - what is it, and what is its prevalence among athletes
  • Risk factors for FLIA
  • Current detection and diagnosis methods of FLIA, as well as treatment options
  • How Jem's research aims to find methods by which FLIA detection can be improved

Sponsored by:


Exceptional quality triathlon wetsuits, trisuits, swimskins, goggles, performance sunglasses as well as prescription eyeglasses and sunglasses.  Online vision test for prescription updates and home try-on options available for eyeglasses. Ships from  the US, UK and EU. Trusted by world-leading athletes such as  Lucy Charles-Barclay, Javier Gómez Noya, Flora Duffy, Morgan Pearson, Summer Rappaport and others in triathlon, cycling, speed skating, and many more. Visit for 20% off your order.

Zen8 Swim Trainer Logo

The ZEN8 Indoor Swim Trainer allows you to improve technique, power, and swim training consistency. You can target specific aspects of your stroke, like catch and pull-through, work on core activation and body position, and make sure you stay consistent in your swim training even when you don't have time to go to the pool. Try the Zen8 risk-free for 30 days, and get 20% off your first order on


Jem's background

04:45 -

  • I am a PhD student at the University of British Columbia in Vancouver, Canada. 
  • I am currently living and working in the Netherlands.
  • I am a physiotherapist, exercise physiologist, and coach of different sports.
  • Much of my work is done with NIRS (Near-Infrared Spectroscopy), doing metabolic and physiological testing for athletes and in a research context.
  • I am studying clinical conditions. However, that affects elite, highly trained endurance athletes.

What is NIRS

06:10 -

  • NIRS is about shining a light into the tissue, allowing for non-invasive analysis like a watch. The light bounces around that translucid tissue, and some light will return to the sensor.
  • We primarily focus on specific wavelengths: wavelengths that oxygen-carrying molecules absorb. (haemoglobin and myoglobin)
  • The signal will tell us something about how much oxygenated those "heme cells" are and how much they are lacking.
  • Let's look at a muscle engaged in an activity. We will be able to see something about the metabolic activity in a little area of tissue we illuminate with the sensor. (SmO2. Muscle oxygen saturation - per cent saturation of the "heme" cells)
  • Depending on the person, it will start at 60 %, and it can go up to 90 % and down to close to 0 %.
  • Several studies tried to separate haemoglobin and myoglobin, but we cannot differentiate the two for the most current and portable devices.
  • However, we can say myoglobin does not change and increase with exercise. So, we can be confident that myoglobin stays constant and that any change we see during exercise comes from the haemoglobin side. (changes in blood volume and blood flow in that area of illuminated tissue)
  • As the "heme signal increases, haemoglobin increases by an increase in blood volume or an increase in hematocrit in the local capillaries.
  • SmO2 looks at the mixed capillary blood, meaning the arterial side with high oxygenation (99-98 %). Along that single capillary, oxygen will go to the muscle, and the saturation can be very low on the venus side. (oxygen extraction can be 90 %, and that is why we get an extensive range of SmO2)
  • Generally, we measure the mixed venous blood, meaning we will never see 0 % saturation. Even at the venous part of the capillary, we will not have 100 % of oxygen extracted from that capillary.
  • SmO2 measures predominantly on the venous side, but that will change as the tissue changes with activity (dilatation of capillaries and the mechanical forces on the tissues).

Accuracy of different devices

13:03 -

  • Most sensors used are little portable units which are self-contained.
  • There are two sensors I have used the most.
  • There is much validation for Moxy, and one of my colleagues did many experiments looking at the reliability of Moxy in real-world conditions. (in a training environment to understand the daily variation)
  • Typically, there is a 10 % difference in SmO2. If in one day you are at 50 % and in another you are at 60 %, and we will assume no significant difference.
  • As you use it more for a single individual, you narrow that range to the individual.
  • The other sensor I am using is the PortaMon, one startup within our nearest manufacturers. We measure data at 100 Hz, but it does not have that validation and reliability study.
  • The precision is around 5 %, but there is a difference in the actual percentages you will get.
  • For Moxy, you get the 0-100 % range, and for PortaMon, 20-80 % range.
  • Most early literature uses clinical equipment, where the exact numbers differ from portable devices.
  • We presented the accuracy levels from Moxy from data obtained during severe intensity domain exercise.
  • If we do a ramp test, there are differences in accuracy from low to high intensity.
  • However, data variability at a high intensity (10-15 %) will always be higher than at low intensity (5-10%). 

Using a SmO2 monitor for endurance training

20:02 -

  • The signal will depend on the protocol (variable vs constant workload).
  • At rest, most people will be around the same values. (50-60 % using the Moxy)
  • It will depend on tissue composition and pigmentation.
  • We see SmO2 increase at a low intensity because of the warm-up effect. (there is an increase in oxygen delivery to the tissue relative to our constant workload. The delivery is higher than the demand, so the SmO2 trends upwards to 80-90 %)
  • The signal has to go through the skin and fatty tissue before getting to the muscle. That tissue is less metabolically active and contributes to the signal received.
  • When we are warming up, there will be an increase in cutaneous blood flow for thermoregulation. We do not see the skin extracting much oxygen, increasing the saturation signal.
  • If we keep that low intensity for an extended period, SmO2 will stay stable. It slightly decreases over time, but it stays high.
  • We will have a drift just like with HR.
  • SmO2 is the balance between oxygen demand and supply. Just like blood lactate is the net result of blood lactate cleared and produced, SmO2 is the balance between oxygen delivery and extraction in a specific tissue.
  • If delivery is higher than extraction, SmO2 will go up.
  • However, we cannot say the absolute magnitude of oxygen delivery and uptake.
  • For example, SmO2 will increase after an intense bout, meaning more oxygen is delivered to that tissue than oxygen extraction.

SmO2 variation at higher intensities

26:26 -

  • In high-intensity training, there are three phases of the interval: the onset kinetics, where you go from rest to 250-300 W and impose a sudden increase in demand for energy that we try to supply with oxygen, meaning we will see a quick desaturation of SmO2 during the first 1-2 min of effort.
  • It is the inverse of Vo2.
  • Then, we enter a relatively steady state, where SmO2 will tend to balance, achieving a minimum at the maximum exertion, where we continue to deoxygenate until the end of our interval.
  • When we stop exercising, energy is no longer needed. Still, oxygen delivery has ramped up during exercise, oxygen extraction is lower than oxygen delivery, and we have a steep recovery of SmO2.

SmO2 in a progressive step test

29:49 -

  • If we do a lactate test protocol, SmO2 will continue to lower as workload increases. (monotonic response)
  • During each step, you see an increase in SmO2, but as you move to the next stage, there is a more significant drop in SmO2 than the increase during the prior stage.
  • A typical blood lactate test will have 2-3 stages in the moderate intensity domain. The change in workload between steps causes higher deoxygenation than you get in stage one. In each stage, there is an upgrade slope in SmO2 that is not very steep.
  • The increase in workload in the next stage leads to an increase in oxygen demand, and you will see this monotonic profile.
  • In the heavy domain, the deoxygenation increases between stages and the slope within those stages will be flat/slightly negative.
  • When athletes reach critical power (MLSS), they will be near their physiological SmO2 minimum, meaning they cannot deoxygenate further.
  • Therefore, we enter a second part of the curve in the severe intensity domain.

Second Response curve profile

34:29 -

  • The second response curve language is: curvilinear, non-monotonic or parabolic.
  • At the moderate domain, we see the same response with the workload, but the increase during the stage is enough so that at the next stage, the SmO2 is higher. (Higher at stage 2-3 compared to stage 1)
  • You reach the maximum SmO2 at the VT1 or LT1. 
  • In the heavy domain, it is more of a plateau at the maximum SmO2.
  • In the severe intense domain, we will see steep, sharp deoxygenation to a minimum at task failure.
  • Typically, leaner male, fitter athletes present the first response profile. The second profile is for less fit and less lean athletes and more for women.
  • There might not be a sex difference, primarily because of the tissue composition. Even fit women athletes will have more fat in their tights than male athletes. It means women will have more blood flowing through non-active tissue (skin) than male.

Implications of having different SmO2 response curves

37:40 -

  • Suppose I do a session where I want SmO2 and energy substrate to be maximal, with the first athlete with a monotonic response. In that case, the workload will tend to zero, and they will work at a significant different workload than the second athlete.
  • For high-intensity sessions, the curvilinear response curve athletes, SmO2 will tend to a minimum, so we can modulate the intensity and duration to keep that slope to avoid hitting task failure before the end of the effort.
  • The linear response athlete is already at a minimum at critical power, so there is no slope to predict the time-to-exhaustion of those athletes, and we cannot use this metric to program high-intensity intervals.

SmO2 testing protocol

41:09 -

  • We need to understand the context where we measure this parameter to understand the response. Most literature uses an 8-12 minute standard ramp test.
  • With NIRS, we want to look at the steady state response. In a continuous ramp test, we only see onset kinetics, meaning we never achieve a steady state.
  • So, we have a multi-stage incremental test just like lactate testing.
  • We introduce a one-minute rest between steps to show us the oxygenation response from a prior workload. (an approximation for oxygen delivery)
  • Therefore, we can understand how the onset and recovery process progresses.
  • We have been using this protocol for zone training description, but we are still validating the results and the athlete's responses.
  • We have some work comparing NIRS with other testing protocols and some reliability studies where we test athletes twice.

NIRS and VO2 relationship

44:51 -

  • NIRS is less reliable than VO2, while NIRS might be as reliable as blood lactate.
  • The problem with lactate is that you only get one data point for each stage.
  • It means that if you train with a sensor for two consecutive days, you have to understand the uncertainties of the different devices.
  • It might not be meaningful if you see a 15 % difference at a high intensity.
  • So, it is where the athlete and coach will study and understand if those differences are significant or not.

Relationship between SmO2 and physiological parameters

47:01 -

  • One application of the SmO2 is the determination of the deoxygenation breakpoint (deoxygenation plateau/ DHB plateau). This breakpoint relates to critical power/VT2/LT2.
  • We also have an oxygenation breakpoint, which may correlate with the LT1/VT1.
  • However, these points depend on the oxygenation profile: with a curvilinear profile, the SmO2 maximum relates to VT1, but with the monotonic profile, there is no association between VT1 and SmO2.
  • In a study, we found a correlation between the deoxygenation breakpoint and VT2. However, this plateau can be a plateau or a more steep decline.
  • We place the sensor at the deltoid. (a non-locomotor muscle)
  • You will lean more on the bars at higher intensity, using more of your upper body to stabilise your torso.
  • There will be some extra work that will drive the increase in oxygen extraction.
  • Some studies showed that you still saw the same deoxygenation response even with no extra muscle recruitment.
  • If there is no increase in recruitment and SmO2 is going down, oxygen delivery lowers.
  • The systemic vasoconstriction at higher intensities makes oxygen extraction exceed oxygen delivery, so we see a deoxygenation response.
  • In our study, the group mean showed no difference between the different physiological breakpoints, but there is a high degree of variability at an individual level. For some athletes, VT1 occurs, and then we see SmO2, and for others, the opposite.
  • We did not see a consistent order of the different breakpoints, suggesting these breakpoints cannot be used anonymously.
  • Therefore, the prescription has to consider these uncertainties, but you can get to a high level of precision individually.

Applying NIRS to an athlete

54:51 -

  • It is easier to implement high-intensity training because, for the second profile (rapid deoxygenation profile), you modulate the power/duration of each session.
  • If we see an athlete falling in this curvilinear profile in the baseline test, we might send a couple of sensors for the athlete to use for some weeks. The goal is to see what SmO2 looks like during training. We can test different intensities and figure out how the deoxygenation rate varies. (trying to triangulate Power, HR and NIRS)
  • For low intensity, it is all about sensations. However, having NIRS in real-time allows us to evaluate the impact of short climbs and dial the sensations, power and heart rate.
  • After a few weeks, they return the sensor but remember the sensations and know what intensity they work.
  • Moxy has many resources that evaluate different athletes' potential strengths and limitations.
  • We used NIRS to understand the relative limitations and strengths. And we can reach a point where data does not seem to make sense, however, if we use other metrics to evaluate other parameters around the system.
  • Probably, athletes with a monotonic response curve cannot extract more oxygen in that area of tissue (there is no more oxygen left)
  • The other profile that continues to deoxygenate further to task failure seems to have more delivery. If those athletes work on oxygen extraction, they might have that SmO2 lower and extend that time to exhaustion.
  • We can also get more information about the athlete's phenotype and what training this athlete should be doing to optimise performance.
  • If you have someone new to triathlon and have more experience in cycling/running, the responses will be different across modalities, and NIRS can be different.
  • While we cannot take "cycling numbers" and apply them to running, we can translate them between sports if we can dial in the sensation.

Workouts for different types of athletes

1:06:13 -

  • If we see an athlete at a minimum SmO2 at critical power, we can work "as hard as possible" while maintaining that SmO2 above that limit.
  • Once that SmO2 is at a minimum, the workload increases from anaerobic substrates.
  • Keeping that oxygenation above the minimum and getting that athlete to work for an extended period where VO2 is high will allow us to improve.
  • The intensity in itself does not matter as much.
  • We can also do sprint interval training (e.g. 20s all-out efforts where oxygenation will plummet) for athletes with a curvilinear response.
  • They might not be able to get SmO2 low, so we want to expose the tissue to that stimulus.

Devices for working with NIRS

1:10:10 -

  • The most well-known devices are Moxy and TrainRed (previously PortaMon).
  • Moxy has been around for many years and has many resources, validation studies, and many projects have used it (and it is waterproof so that you can use it in the swim).
  • A study used PortaMon on a pool, but with a custom "protection."
  • These devices can be around mid 600-800$.

Who should buy a NIRS device

1:12:58 -

  • Some athletes have the time to understand the data, turn it into information and apply it to their training.
  • However, not everyone has the time and inclination to do it.
  • So, I imagine a coach having many different sensors, and the athletes come in and can establish their profile.
  • Then, we give the athlete 2-3 sensors for a few weeks to gather data.
  • During this period, athletes can dial their sensations with the NIRS data and connect with parameters like HR and power.
  • Then, coaches can look at that data and find some exciting information.
  • However, I do not think it is for all athletes.

Iliac artery flow limitation

1:15:26 -

  • My PhD is on this topic, and the other metabolic work is understanding blood flow limitations.
  • It is a condition where the primary artery feeding your leg cannot give as much blood flow as required.
  • Delivery cannot match the demand in the legs (it is typically the left leg).
  • The limitation is at the hip/pelvis. 
  • During an exercise bout, if blood flow cannot increase, the tissue downstream becomes hypoxic much earlier and more severely.
  • These burn sensations happen at a lower intensity and are progressive over time.
  • Typically, it relates to the body position on the bike where the knee is at the top of the hip (tight hip flexion). It could potentially put a bend and a kink in the artery.
  • Not everyone is susceptible to this. If you are a time-trialists does not make you susceptible to developing FLIA (Flow limitation).
  • We do not have the exact data, but as many as 20 % of professional cyclists will develop this at some point.
  • It is related to hip flexion and artery forces applied from the surroundings. Psoas muscle recruitment can put pressure on the artery, and in a highly trained athlete, you will have a lot of blood flow trying to go through this "kink".
  • These hemodynamic forces cause turbulence, and the arterial walls do not like it because they can become dysfunctional and start fibrosis, which is to get thicker.
  • It can start as an artery bend and progress to endo-fibrosis.
  • In the early stages, it is challenging to diagnose.
  • As athletes go through all treatments, and the legs get worse, weaker and feel like they start pedalling "squares", it is a potential indication of a flow limitation.

Goals of Jem's PhD

1:20:20 -

  • It has clinical and performance goals. We want to detect this condition earlier and send the athletes to diagnosis.
  • Then, we want to understand the physiological response of the condition.
  • The work we are doing with NIRS (Ramp tests), combined with VO2 and high precision pedal power meters, aims to understand the response across intensity in healthy and flow-limited athletes.
  • We want to differentiate those responses to use them to detect this problem at an earlier stage outside of the cardiovascular specialist office.
  • The diagnosis will continue to be a highly-specialised process. However, if athletes come to a physiologist for a screening test, we might see an imbalance and send them to further diagnosis.
  • There is also the physiotherapist aspect because there are not many treatment options.
  • Some World-class athletes undergo surgery at the end of the season because there are no other treatment options. (bike fit, training, physiotherapy to prevent athletes from going through surgery)
  • There are many case studies about patients who underwent surgery and had a follow-up for 2-3 years. The long-term outcomes are predominantly positive.
  • However, there is no other option, so we want to investigate and create alternative treatments. Professional athletes have no option other than going through surgery because the alternative is to finish their careers.

Detection procedures

1:24:45 -

  • Unfortunately, I have this condition motivated me to pursue this research.
  • I have much anecdotal data on myself. There is variability in this condition for different athletes.
  • However, working with other athletes allowed us to evaluate if the detection methods worked or not.
  • As we talk about a chronic problem, you develop compensatory responses over the years. It is not only a leg imbalance or leg oxygenation in one leg. 
  • The first step will be raising awareness of athletes' potential diagnoses.

Other considerations about FLIA

1:26:40 -

  • The symptoms are burning in the quadricep/calf/glutes coming at a lower intensity than you would expect in one leg.
  • The symptoms also go away quickly after rest. 
  • We have a self-questionnaire (in the links below) describing the symptoms. If this sounds familiar, athletes should go deeper and understand if they are in line to develop this problem.

Blood-occlusion training

1:28:35 -

  • I have not applied this methodology. I understand the theory, and it seems to be promising. I am slightly hesitant because some considerations about automatic parasympathetic hyperexcitability with vascular occlusions exist.
  • There are suitable applications out there. If you want to feel what FLIA feels, do this type of training.

Rapid-fire questions

1:30:22 -

What is your favourite book, blog or resource?

I will put them in general as podcasts. There is no better way of listening to experts talking about different topics.

What is an important habit that benefited athletically, professionally or personally?

It is Consistency and showing up in every part of life. We can add so much data, but adding data can show us the variability in data.

Who is someone you have looked up to or who has inspired you?

Clinician scientists (coaches that work and research). The research and application of methodologies are not mutually exclusive, so I find impressive people who can balance both areas and be successful.


Bernardo Gonçalves

Bernardo is a Portuguese elite cyclist and co-founder of SpeedEdge Performance, a company focused on optimising cycling and triathlon performance. He writes the shownotes for That Triathlon Show, and also produces social media content for each new episode.

{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}

Explore our products and services