Pattern Analysis

This page is in development. 

1. Basic Pattern Analysis for Blood Glucose Monitoring

Request that patients record blood glucose results on paper, in a log book, or use an app/software program associated with the glucose meter. Explain that it is most helpful to see results organized by time of day to assess for patterns.

  1. First, look for pattern of low blood sugars.
    -consider asking, "How often are your readings below 4 mmol/L? How often are you are weak, shaky, sweaty?"
  2. Second, look for pattern of high blood sugars.
    -not all readings need to be in target - don't worry about the occasional, higher readings
  3. Ask yourself if A1c level agrees with results?
    -if not, consider whether A1c is accurate or due for repeat
    -if not, consider that you may be missing data that could be contributing to higher/lower A1c level


2. Basic Pattern Analysis for Continuous Glucose Monitoring using Ambulatory Glucose Profile (AGP)

 As most CGM systems are collecting glucose data every 5 minutes, it is helpful to look at the large amount of data in a combined, graph form. The Ambulatory Glucose Profile (AGP) is a common tool used by all CGM systems. The AGP combines glucose results over several days or weeks into a single graph.

The following is an example of an AGP.

Basic Diabetes Workshop participants: The following is Darpak's AGP results for the past month.



First, understand the graph:

  • Target Range:
    • Appears as light grey band between the low and high targets.
    • Usually comes preset into the CGM device and may need to be changed to reflect patient’s A1c target.
    • Remember to set higher target for after meal excursions – for example, set up to 8 or 10.
  • Orange line- #6:
    • Represents the median – half glucose levels are above this line and half fall below.
    • Ideally, mostly flat and within target range.
  • Blue shaded area -#7 (interquartile range or blue river):
    • Represents 50% of all glucose levels.
    • Ideally, the space narrow and within the target range.
    • The wider the space the greater the glycemic variability.
  • Area between dotted lines -#8:
    • Represents 80% of all glucose values.
    • 10% of glucose values are above the top line (90th percentile) and 10% of glucose values are below the bottom line (10th percentile).

Then, follow same steps as for Blood Glucose Pattern Analysis:

  1. First, look for pattern of lows.
    -Does any part of the curve fall below the low target? What time of day is this happening?

  2. Second, look for pattern of highs.
    -Does any part of the curve go above the high target? What time of day is this happening?

  3. Lastly, does the A1c level agree with the results?
    -If not, consider whether A1c is accurate or due for repeat
    -If not, consider whether you are missing data that may be contributing to higher/lower A1c level.





The Myth of the Somogyi Effect

The Somogyi Effect refers to high fasting glucose readings caused by counter-regulatory hormone release from untreated overnight hypoglycaemia. Despite most evidence failing to support this effect, the concept of the Somogyi effect persists in many health care providers and patients, preventing them from increasing overnight insulin.

Choudhary et al analyzed patient data and concluded the Somogyi effect is indeed, very rare. Their research confirmed previous findings in patients with Type 1 diabetes that overnight hypoglycaemia is associated with low fasting glucose readings.

Taylor outlines some history of how Michael Somogyi (a biochemist) became associated with morning hyperglycaemia secondary overnight hypoglycaemia (even though morning hyperglycemia was not part of his hypothesis and not mentioned in his summary.) Observations that Somogyi made in 1938 and published in 1959 were that five young people had markedly unstable blood glucose control with frequent hypos and periods of excess glycosuria; control was re-established by decreasing insulin doses. Excess secretion of the counter- regulatory hormones was postulated as a cause. Taylor notes a more likely cause of hypoglcyemia in the subjects was the transitioning from the insulin resistance of adolescence, which required dose reductions.


Choudhary, P et al. Do high fasting glucose levels suggest nocturnal hypoglycaemia? The Somogyi effect-more fiction than fact? Diabet Med 2013; 30: 914–917. DOI: 10.1111/dme.12175

Taylor, R. Things that go bump in the night Diabet Med. 2013 Aug;30(8):889-90. DOI: 10.1111/dme.12183.