We often get asked questions about whether sensors can be worn in different spots or whether the sensor would be more accurate if worn on a working muscle.
Until recently, we were bound by the limitations of what had been developed by Abbott and some scientific reasoning but thankfully now researchers have done the work for us in answering some of these questions.
In a paper released by Coates and colleagues using Supersapiens, the researchers tested different placements of the sensor in response to different stimuli. Specifically, arm and leg placement (Tricep and Quadricep) after ingesting 100g of glucose in four conditions; rest, graded cycling, electrically stimulated quadricep contraction and heating without movement.
That’s right, participants wore 2 sensors and compared the sensors to finger capillary glucose.
This study design is elegant in that it attempts to not only answer the question of whether there is a difference between sites, but also what may drive the difference.
What did the Researchers Find?
- At rest the tricep sensor position was closer to the finger capillary glucose.
- Exercising (cycling specifically) narrowed the discrepancy between the leg sensor and the arm sensor as well as finger capillary glucose.
- The researchers suggest that there may be a role for wearing the sensor over exercising muscle.
Some Factors to Note and Consider:
1. Reference values
The reference values used by the scientists were finger capillary glucose. As discussed by the authors of the paper and previously, as well as on our podcast this may not be of significant relevance given interstitial and blood values are measuring different things (and interstitial glucose may be more relevant in athletes). Similarly, there is an acceptable error (15%) for glucometers (the devices used to measure finger capillary glucose) too, and unfortunately not all on the market meet these requirements. (Note we are not saying this is the case with the glucometer used in this case, but more generally for informational purposes)
2. Blood Flow
The insights from this paper make intrinsic sense when looking at things from a purely physiological standpoint (though this should never be the only lens we view technology through). Exercising will increase vasodilation, and thus blood flow peripherally. Which in turn will increase the potential for glucose to perfuse out of the blood into the interstitial fluid (because diffusion pressure is impacted by concentration gradient and flow rate) as is corroborated in this paper.
3. Muscle Fiber Type
Generally, glucose utilization is higher in muscle with a greater proportion of more ‘fast-twitch’ type muscle fibers. This is because their oxidative capacity is lower. As such, it stands to reason that glucose concentrations may be slightly higher here when all else is equal to ensure demand is met if needed.
4. Sample Size
The nature of science is such that scientists are always ‘sampling’ a population. Ideally this sample is representative of the population, including being big enough such that any statistical noise is drowned out. In this case, the nature of this initial exploratory study is such that the sample size is quite small. This increases the chance of some of the results not being reproducible (read ‘real’ or ‘true’). This should not be construed as criticism, it is advice for readers to ensure these results are interpreted in context. In fact, the most reasonable hypothesis following this study is that these results are in fact true, not the opposite.
Supersapiens’ Take:
- The challenge with using a sensor on a specific body region is which to choose. Whilst the vastus medialis position (inner quadricep) makes some sense, this may not be the most relevant to sports other than cycling (or even in cycling for that matter).
- Similarly, given the tricep placement being more accurate for non-exercise time it could be argued this is more valuable given the proportion of exercise vs non-exercise time (even in top endurance athletes).
- In addition to all this, the current limited understanding of glucose kinematics in individuals without diabetes, using CGM technology during exercise means choosing a specific area may introduce even more variability in values for interpretation and understanding.
- From a practical standpoint sensor placement on the leg is inconvenient and hazardous for extended sensor life.
Take Home Message:
- This sort of research is vitally important in the understanding of CGM technology in people without diabetes and Supersapiens is extremely proud to be able to help researchers in this pursuit.
- Sensor site likely does not impact sensor accuracy meaningfully.
- The approved sensor application site is the tricep, despite some users using other locations.
A special thanks from the Supersapiens team is extended to all researchers and subjects involved in this study. Similarly thanks are extended to the publishing journal and reviewers.
References:
- Alexandra M. Coates, Jeremy N. Cohen & Jamie F. Burr (2023) Investigating sensor location on the effectiveness of continuous glucose monitoring during exercise in a non-diabetic population, European Journal of Sport Science, DOI: 10.1080/17461391.2023.2174452
- Siegmund T, Heinemann L, Kolassa R, Thomas A. Discrepancies Between Blood Glucose and Interstitial Glucose-Technological Artifacts or Physiology: Implications for Selection of the Appropriate Therapeutic Target. J Diabetes Sci Technol. 2017 Jul;11(4):766-772. doi: 10.1177/1932296817699637. Epub 2017 Mar 21. PMID: 28322063; PMCID: PMC5588840.
- David C. Klonoff, Joan Lee Parkes, Boris P. Kovatchev, David Kerr, Wendy C. Bevier, Ronald L. Brazg, Mark Christiansen, Timothy S. Bailey, James H. Nichols, Michael A. Kohn; Investigation of the Accuracy of 18 Marketed Blood Glucose Monitors. Diabetes Care 1 August 2018; 41 (8): 1681–1688. https://doi.org/10.2337/dc17-1960
- MacLean DA, Bangsbo J, Saltin B. Muscle interstitial glucose and lactate levels during dynamic exercise in humans determined by microdialysis. J Appl Physiol (1985). 1999 Oct;87(4):1483-90. doi: 10.1152/jappl.1999.87.4.1483. PMID: 10517782.
