My Blood Glucose Data

Context

A friend of mine gave me a continuous glucose monitor (CGM) - specifically the Freestyle Libre 14 Day System. For 10 days, I wore the CGM and concurrently tracked everything I ate and the physical activity I did. That included surf sessions, yoga classes, cold baths, jacuzzi sessions, and hikes. As a note, I am not diabetic and used the device out of curiosity and potential health optimization.

Before jumping into the data, here is a little background about blood glucose. First, glucose is important to our bodies because our individual cells use it for energy. Normal blood glucose levels (AKA blood sugar concentration) is between 72-99 mg/dL while fasting and 100-140 up to 2 hours after consuming food. High blood sugar (Hyperglycemia) and low blood sugar (Hypoglycemia) are the resulting conditions if someone is consistently outside these accepted ranges. In general, there are many things that can cause blood glucose concentrations to fluctuate. A few include Food intake, Exercise, Hot or Cold Exposure, and Stress. A healthy body keeps blood glucose within a narrow range and does so by releasing catabolic hormones (i.e. glucagon or cortisol) or anabolic hormones (i.e., insulin) to increase or decrease the blood glucose concentration, respectively. 

There has been an increase in the blood glucose monitoring market, and I believe it’s due to a few reasons. 

• Blood glucose monitoring technology is advancing

• Studies have recently illustrated that blood glucose fluctuations vary widely from person to person. This means that understanding our own body’s blood glucose can provide insight into the benefits or harm that food and activities may be causing.

• Additional studies have linked large, frequent fluctuations in blood sugar with increased risks of all-cause mortality. This means that steady and well-regulated blood glucose levels could lead to healthier lives and using a CGM is the first step.

Methods

I downloaded a .csv file of my glucose data and used Python & Adobe Illustrator to produce a few plots. Most data points are 15 minutes apart, but some intervals are larger because I failed to scan my CGM within 8 hours (the Freestyle Libre only saves 8 hours worth of data before overwriting)

Figure 2: Daily plots

In these graphs, I was mainly concerned with the spikes that I identified as abnormal (orange line). I added food consumption (red) and the major physical activities I performed each day (yellow highlight) to help identify the cause of the blood sugar spikes. Note that I did not include everything I ate (because it was clutter the graph) and I only displayed days within the dataset that I found most interesting.

What I learned

• Blood glucose varies greatly throughout the day
  To my surprise, my blood glucose spiked many more times throughout the day than I expected. I believe that I eat a relatively healthy diet, abiding by the 80/20 rule when it comes to healthy vs non-healthy items. During this period of time, I was away from home and ate at more restaurants and more frequently than usual. I am speculating that these factors may have caused the higher and more frequent spikes in blood glucose concentration.

• Blood glucose increases during exercise
  Surprisingly, during every surf session I recorded, my blood glucose increased. This makes sense because during exercise our body needs energy to perform physical activity. When our cells need energy, our body sends a signal to convert our glucose stores (in the form of glycogen found within our liver and muscles) to usable glucose and deposit it within the bloodstream. This leads to increased blood glucose concentration. While physical activity can increase blood sugar in short bursts, over a longer time frame, physical exercise has been shown to keep blood glucose levels lower.

• Gluten and refined sugar triggered the largest spikes in my blood glucose
  Two events towards the end of the dataset triggered a high blood glucose concentration (above 128 mg/dL). The main similarity between these two events was a high carbohydrate load that consisted of gluten and refined sugar. The first event was on October 9th and was followed by the consumption of a sweet taro bun and curry. The items that preceded the second event (Oct 11th) included a burrito and some berry cobbler. During the period when my blood glucose was above 120 mg/dL (50 min), I was in a jacuzzi.  The unhealthy food and a hot environment were the perfect combination to keep my glucose levels raised for almost an hour.

• Food had a latency period between consumption and blood glucose increase
  Depending on what I ate and how long it took me to eat, the time it took to increase my blood sugar would change, but generally, it took about 15-25 minutes between food intake and my body’s absorption/conversion of the sugars within the food. 

• A meal’s macronutrient proportions affected my glucose response
  The ratio of macronutrients (fat, carb, protein) in a meal impacted both the duration and value of my blood glucose. For example, if I consumed just one piece of bread vs one piece of bread with almond butter, the bread with almond butter will offset the following increase in blood glucose. In the plots above, on Oct 6th, I consumed kombucha and watermelon following a yoga class, and in both cases, very sharp spikes followed. These foods were predominantly carbohydrates. Earlier in the day, I ate 2 eggs, 1 piece of toast, and a banana with peanut butter. The carbs (toast & banana) paired with fat and protein (peanut butter and eggs) led to a slower increase in blood glucose lower local maximum.

Conclusion & Future

Overall, the analysis provided a greater awareness of my body’s continual regulation of blood glucose. If my goal is to steady my blood glucose further, I will think twice before consuming any item that is solely carbohydrates, especially if it’s paired with any other factor that may increase my blood glucose.

As a note, in this analysis, I did not spend any time understanding low glucose events but would like to understand these events further. 

Lastly, I plan to use a CGM again to collect more experimental data. I would like to know how my blood glucose appears on a Keto/lower-carb diet and potentially during longer periods of fasting.