Compression Lows

This may be a new opening of an old wound. I have seen and heard of COMPRESSION LOWS rearing its ugly head this fall. I am trying to wrap my head around the term and what is happening in the body. Does this seem like a good description?

BACKGROUND: Compression lows are caused by the person’s body pressing against the mattress, pinning the CGM sensor/transmitter (CGM) between the person’s body and the mattress during sleep. First, the interstitial fluid (IF) is the fluid around body cells. Most of the time IF is exchanged with fluids in the blood vessels. During this exchange, IF glucose is enriched and cellular waste products are removed from the IF all over the body.

THE PROBLEM: When a person is in sound sleep and is mashing the CGM into the mattress, the IF is also mashed. It is this mashing or compressing of the IF that causes the IF not to exchange as it normally does. Since the cells around the CGM sensor wire continue to consume (eat) the glucose available in the compressed area around the sensor wire the glucose in the area drops.

OBSERVATION: It is this low reading in the CGM data is usually seen as a nearly level data graph and then a sudden drop of the glucose level. Because it is a low brought about by the pressure on the CGM, the term COMPRESSION LOW has been offered in pumping circles. Because the Compression Low is a low glucose only in the area of the CGM, a finger stick performed in response to a LOW alarm will show NORMAL. The variation between the finger stick and the CGM data leads to frustration and bewilderment, believing technology is the source of the error.

The next observation is the blood sugar measured by the CGM will return to a value near where the CGM line was before the drop in CGM value. The near level line, the drop & alarm, the finger stick of different reading, and the return to the near pre-alarm value is the full picture of the COMPRESSION LOW.

I would really like to get feedback. Thanks all.

The scientific papers I’ve read mention “reduced glucose diffusion”, which I think matches your theory, but also indicate that compression effects on oxygen concentrations and other interstitial chemistry as well as temperature. Quoting , “Susceptibility of Interstitial Continuous Glucose Monitor Performance to Sleeping Position”:

“What are the potential mechanisms of sensor inaccuracies during sleep? Laying on subcutaneous CGMs may cause pressure-induced or temperature-induced physiological fluctuations. Interstitial glucose sensors rely on adequate blood flow to the interstitial space for accurate measurement. Further, interstitial glucose sensors are sensitive to oxygen tension in the interstitial space; fluctuations in oxygen tension can affect the accuracy of the sensor.32 Finally, in vivo testing has also demonstrated that temperature has a profound impact on both subcutaneous oxygen tension and sensor output current.33,34 Pressure applied to tissue in the immediate vicinity of a sensor may decrease blood flow and oxygen tension and increase temperature, resulting in the CGM inaccuracies observed in the present study. Specifically, decreased blood flow to the capillaries immediately surrounding the sensing electrode is expected to depress CGM readings because of reduced glucose diffusion into the sensing area. Decreased oxygen availability in the tissue will also depress CGM readings due to the oxygen dependency of the enzymatic conversion of glucose via glucose oxidase. Warming of the area due to laying on the CGM will likely have the opposite effect and increase sensor signals. Clearly, CGM readings are influenced by multiple factors, including glucose and oxygen availability from the nearby vasculature, temperature, and other potential unidentified factors (e.g., temporary release of interfering substances during localized pressure or temperature, direct mechanical pressure on membrane elongation, or tension).”


I have not had any compression lows in the past few years and in order to make that happen, I just make sure that the sensor with transmitter is never in a position to get buried in the mattress. Each person sleeps differently. I am a side sleeper, one side or the other so by placing the sensor and transmitter mostly inline between belly button and xiphoid process, the sensor and transmitter is virtually always in direct line of sight to the receiver with no mattress interference.

1 Like

Thanks, Tim.


I think that @Tim12‘s reference makes the important point that the compression reduces the blood flow locally … being more important than “squishing” the IF itself.

I have wondered whether one could make a BG measurement just adjacent to the sensor during a compression low and get a nice, real datapoint to test this theory. I rarely get a compression low, but maybe it would be interesting to intentionally induce a compression low to test this out.

Stay safe!


1 Like

John, If I look at your words, you mention BG, which may mean a “belly” blood draw - is your suggestion that they do a belly blood draw and look and see if it did or did not show the drop seen by the CGM in the ISF?

They do a correlation between CGM’s on same and different sides, but I don’t think this is what you are suggesting.

Sensor excursions were strongly correlated with sleeping on the side of the sensor. For example when not sleeping on the left side , left-sided sensor excursions occurred for 13.4% of time points (50 out of 372; Table 1 , rows 1 and 3), whereas when sleeping on the left side , left-sided sensor excursions occurred for 32.4% of time points (56 out of 173; Table 1, row 2). Using Fisher’s exact test on the contingency table, this correlation is significant ( p < .001). Correspondingly, right-sided sensor excursions were much more common when sleeping on the right side [18 out of 143 (12.6%) time points] than when not sleeping on the right side [25 out of 402 (6.2%) time points; p < .001].

1 Like


Thank you for posting that article and I have now read the whole thing … in addition to your extraction of key parts. Interesting indeed.

To answers your question, yes, I was suggesting taking a blood sample meter measurement adjacent to the sensor at the same time that one was experiencing a compression low. I actually wear my sensor on the back of one arm. While I’ve always done finger sticks, I know that a number of people use the inside of their forearm so was guessing that one could also draw a sample from either belly or upper arm depending on where the sensor was affixed.

I doubt that most lancets would penetrate the sensor adhesive fabric and, even it it did, the blood would like wick into the surrounding fabric instead of into the BG strip. As a result, if I try this experiment, I though I’d wait until a sensor session is about to end, then peel up the adhesive on one side of the patch, then compress the area and try to measure the near-sensor blood glucose at the same time that I’m seeing a compression low.

Of course, reading the paper that you posted is both easier and has better statistics …

Thanks again,


The sensor uses interstitial tissue fluid not blood.
That fluid can become trapped when you lay on one spot for a long time. Or if you are wearing a belt in top of your sensor.
Your blood flows through capillaries, so your blood will not be effected by compression. If it did you would have a much bigger problem on your hands.

Testing near the sensor will not likely tell you anything more than from your finger.

This is a good choice to make. I am also side sleeper and can relate with this.