Dr. E's Case Study

This case is extremely important because we rely on CGM downloads to get an idea of how patients are doing. If you have a time below range of 0%, it would appear that you’re not having any lows. In reality, you could have had a severe low, but it didn’t register on your download because you didn’t have it long enough (depending on the length of the download) and your healthcare provider will miss picking up on the low blood sugar. This is extremely important because questions need to be asked regarding the hypoglycemic event (e.g., what time of day did it happen, was it a compression low, was it due to too much fast-acting insulin, was the basal rate of your pump or your injection too high because you lost weight, etc.). You and your provider really need to be aware that you can miss a hypoglycemic event when looking at these downloads.

The Patient:

José is an 85-year-old Army veteran who has had type 1 diabetes for over 20 years. He’s currently on a multiple daily injection (MDI) regimen, and he has a degree of hypoglycemia unawareness. He came into clinic recently with the 14-day download from his FreeStyle Libre 3  CGM, and at first glance it looked extremely good:

December 2024 Case Study Libre Download

As you can see from the report:

  • His time in range (target range) was 78%
  • His lows/very lows were 0%
  • His average glucose was 146 mg/dL
  • His GMI (sensor-derived A1c) was 6.8%
  • His glucose variability was 31% (goal is less than 36%)

In the AGP chart below, you can see that he had a couple of ups and downs, but with those metrics at 85 years old (and with zero lows) I was very happy.

December 2024 Case Study AGP Report

So…If Everything Looks So Good, What’s the Problem?

After looking at the first-page summary of José’s download, I told him it looked like he was doing great, especially because he wasn’t having any lows. He said, “What do you mean? I have had a couple of bad lows.” I dug deeper into the download and sure enough, he did have a few low blood sugars that were well below 55 mg/dL. You can see the tiny red dips in the glucose pattern chart below.

 

December Case Study Glucose Patterns

Since José’s lows lasted less than 15 minutes in total, they weren’t long enough to register as even 1% on the download (1% represents 15 minutes on average for each day of the download). If your report says you were low 1% of the time, that means you were below 70 mg/dL 15 minutes per day on average for the duration of the download.

The key point of this case is that when you look at the general summary data for a CGM, when it says 0% low or 0% very low, it’s an average of the amount of time you’re below 70 mg/dL and 55 mg/dL respectively for the length of the download.

The Solution:

As you can see from the chart below, the time in range goal for a person with type 1 diabetes who is older with comorbidities is greater than 50% with minimal lows. Although José’s time in range was excellent, as an 85-year-old man, he does not need to be so aggressive with his goals. He also shouldn’t be experiencing any lows at all. When you’re 85 years old, 65 mg/dL is dangerous. I reduced José’s basal insulin by 10%, and I asked him to check in with me in a few weeks.

Consensus on time in range targets for people with diabetes

The Takeaway:

The key point in this case is that José had some pretty bad lows, but a provider wouldn’t have known that from looking at the download summary.

You have to remember that these summaries average out all the data (GMI, glycemic variability, lows, etc.), and if you’re not low for a long enough period of time, the download will read 0%.

As a patient, if your download summary doesn’t reflect your blood sugar history accurately, let your provider know. If you are a provider, look for trouble spots in the entire download, and check in with your patients about any concerning hypoglycemic and hyperglycemic episodes they have experienced.

 

8 Comments
  1. Avatar

    I didn’t think about the fact that a low wouldn’t show up because of the short duration. Thank you for this knowledge.
    I went to donate blood and was told my hemoglobin was too low (10.3). Around the same time my A1C was over 7 even though my CGM analysis suggested it should be around 6.4. I hadn’t realized since the A1C measures hemoglobin that being lower than normal would artificially increase my A1C results. Another reason all data points are important!

  2. Avatar

    Thanks for the details, however wanted to know what was the treatment for Jose to avoid the dangerous lows?

    • Avatar

      Hi Sam,
      For now, Dr. Edelman reduced his basal insulin by 10%, and José is going to check back in after a couple of weeks to see if that helped his lows.

  3. Avatar

    Thank you. I will review past CGM data. I have frequent lows but generally correct them before the 15 minute mark

  4. Avatar

    Hi Dr E –

    I’ve been watching your YouTube channel for a few years now and recently wanted to see what your comments on gastroparesis were.

    My neurologist recently said my tremors were from Autonomic Neuropathy as well (not Parkinson’s)

    I am nearly 70 now with 38 years of diagnosed Type 1

    What is the best and simplest way to control sugars without losing your mind, went off the Omnipod 5 because I couldn’t figure out what it was doing. Currently using InPen with a G7.

    Thanks, Lori

  5. Avatar

    I downloaded the data during a 45 day period when I had lows where the CGM only indicated “LOW”. The summary the NP looked at had 0% very lows for the current 30 days. BUT when I looked at ALL the rows of data it does average out to 0%. I have to take a report with me that indicates how many times I was very low, low, in range, etc
    when you look at 1000s of rows of data – a few very lows doesn’t amount to much

Leave a Reply