Ultrahuman and experiments with Fasted Running

I have been experimenting with "Running on Empty", which loosely defined translates to running on a carb depleted state. I now regularly run all my morning runs on a overnight fast of about 10hours, by which time the blood glucose is almost depleted.

More from curiosity than anything else, I decided to be a Cyborg for 2 weeks - to try the UltraHuman program from the Bangalore based start-up. It uses a sensor patch which detects interstitial fluid to measure glucose via a probe. The patch transmits live glucose data to the "Continuous Glucose Monitoring" (CGM) app.

Day0:

Wearing the sensor. The product comes in a sleek package. During the recommended overnight fast before inserting it, I was quite anxious. Esp. after the instruction video indicated that there could be a "few drops of blood splatter". My heart rate went up as I inserted the plunger into my arm, cringing for pain, but lo! nothing happened. I was a Cyborg!!

After a warm-up, there is an apparent, 24hr calibration period. My sensor went flat during this time (see pic below). The customer service was prompt and shipped a replacement the very day.

I had to peel off the dysfunctional one and do the insertion all over again, minus the anxiety this time. It did take that 24hr to calibrate, see the night readings being haywire on day one below.

It was pretty awesome to see the readings and the variations - preprandial, postprandial and nocturnal - live and on the app. Also, the gamification built into the app, with a KPI of "metabolic score", it kept me checking the app quite often in those days. My glucose range stayed within the recommended range of 70mg/dL to 110mg/dL.

That got a bit boring and I decided to "run" some experiments! These are not your typical "I order XYZ from restaurant ABC on swiggy and my glucose went up to 180..." experiments. 

The guinea pig (read: me) has been on a "fat optimization" journey for the last few years. Bear with me, these were with n=1, age: 41kg; BMI: 21, with no randomized or blind test control sets.

Experiment #1: Tempo paced run

Control variable: 24k at tempo (5min/km) pace 

Treatment1: with '0' calories (empty run)
Treatment2: with carb intake at 5k and 14k during the run

The near linear drop in the glucose level on the top graph is probably an interpolation, but the important thing to note is that the end state glucose level at 8am. In the empty state, it was 60mg/dL, whereas with the gels it was at 70mg/dL. Also notice how the body recovered from the 60mg/dL to 70s without any card ingestion post workout. This probably indicates Gluconeogenesis, where the body is breaking down protein and fat and converting into glucose. Read more on this here -> Ultrahuman Blog 

Result: No Perception of Effort (PE) effort with or without gels

Experiment #2: High Intensity Interval Training (HIIT)

Type: 400m x 20 repeats at 90% of Heart rate
Treatment1: overnight fast and '0' calories during
Treatment2: overnight fast, Gel (110k calories) after 6 intervals

In highly fat adapted athletes, Glucose should rise from stressful workout, without ingestion of carbs. The Gel seemed to have arrested the slide in glucose levels. In the past, I used to feel a serious lack of energy during the last few intervals. But overtime, this feeling has faded a bit - either I am not pushing hard enough OR I am getting better fat adapted (or both!).

Result: The Gel did make a difference, but needs some more experiment to discern if it was the gel or the water that went along with the gel!

Experiment #3: Easy 10k

Type: 10k at 60% - 70% of Heart rate
Treatment1: overnight fast and with '0' calories during
Treatment2: 35hour water-only fast and with '0' calories during

This was interesting. The first 10k after an overnight fast is the regular easy early morning run. As for the second 10k, after the 36hr fast, the first 2km or so felt tired - notice the glucose levels dropping to sub-60 levels. Then it felt good till the end of the run. Also notice how the glucose levels go up to about 70mg/dL post the run. Lipolysis, gluconeogenesis could be either or both in conjugation.

Result: There is a lot of merit in medium intensity exercise combined with fasting, to promote fat metabolism and maybe even accelerate autophagy. Certainly plan to continue this. My best is a 50km run bring up a 36hr water-only fast in 2019.

Experiment #4: Long Slow Distance (LSD) 50k

Type: 50k at 60% - 70% of maximum Heart rate (MHR)
Treatment1: overnight fast and with '0' calories during

Treatment2: increase in pace to about 80% of MHR 3.5hrs into the run for 2km

An early start on a cool day in Bangalore, ran for most of this around Cubbon Park, since it was weekend lockdown with the park closed. Paced a friend, Akhil for a sub-2 hour half marathon. The speeding up was from 35k to 37k, where we went from 5:30pace to about 4:45pace. You can see how this shot up the glucose levels. All this was done on an overnight fast. Initial dip in glucose level to 60mg/dL. This could be indicating fat adaptation. The increase can also be due to cortisol level increase leading to glucose increase. Notice how the increased glucose kept rising, despite no crab intake, even as I resumed running at 5:30 pace from 40k to 50k.

Result: This was one my easiest 50k runs, LSD on empty? Check.

Experiment #5: Typical day

Type: No exercise
Treatment1: No exercise, typical regular home-cooked food

Treatment2: Full day, water-only fast

Fasting has lots of benefits to keep the metabolic score (a term UltraHuman uses generously (Read details here). It keeps all the 3 vectors - glucose variability, average glucose and time in target zone, within the limits.

In closing, even without having to go through these "experiments", the UltraHuman CGM is certainly a great biohack to get you good insights to how your glucose levels react to your typical day. I do wish they include lactate and ketone level monitoring in addition to glucose to given a wholistic perspective of stress & fueling.

Call to Action: If you have come this far into this post, follow me on instagram @ultrabhat (link here)!!!