My first prolonged water fast

January 24, 2021

A cartoon graphic of autophagy

Autophagy, Greek for "self-eating", is a process by which our bodies recycle cellular components.

B and I recently completed a prolonged, water and mineral-only fast. People seemed curious, so I wrote up some notes.

Why would you do this?

Short answer: longevity. Well, maybe.

Do we know for sure that fasting will improve longevity in humans? Nope.

Outside of stuff like regular exercise, not smoking, not drinking alcohol in excess, and reducing sugar consumption, scientists have yet to reach consensus on therapies for postponing the onset of age-related diseases like cancer, cardiovascular disease, and neurological diseases.

Studies that examine disease prevention in humans take a really long time, making them costly and often infeasible. That said, there’s evidence in animal models that fasting may help reduce mortality from many chronic diseases.

Some scientists think that autophagy, a program built into our genetic code for recycling dysfunctional cellular components, may be the mechanism that explains why fasting causes animals to live longer. Autophagy is happening all the time in our bodies at low levels, but nutrient starvation dramatically upregulates it.

I don’t have 40 years to find out whether I should be fasting, so I’m giving it a shot despite the limited evidence. And if it turns out there’s no benefit? Then I didn’t eat for three days. Oh well.

Performance benefit?

I don’t think so there’s a good case that prolonged fasting would do much for athletic performance. If done for too long or with underweight individuals, it could be harmful.

Some people claim that fasting allows for faster recovery, perhaps facilitating injury prevention, but I think there are better, less uncomfortable strategies.

If there is a performance effect, I think we’re talking about the tiniest of marginal gains, and I wouldn’t hold my breath for anything noticeable, positive or negative.

Why not “intermittently fast” (IF)?

IF sounds convenient - “Just stop eating by 4pm every day and you can prevent cancer!”

But the evidence that IF does anything significant in humans for lifespan is, well, lacking.

Mice seem to exhibit upregulated autophagy after about 24 hours of fasting. Does that mean that restricting one’s diet to one meal per day should upregulate autophagy? We don’t know, but my guess is probably not.

Fasted mice can lose about 10-20% of their body weight in a day while humans lose closer to half a percent. I don’t think it’s a leap to think it takes humans longer than mice to upregulate autophagy.

IF may very well have health benefits, and it is more sustainable than prolonged fasting, but if the goal is autophagy, something more restrictive than 16/8 is probably necessary. IMO, IF is likely a complement - not a replacement - to prolonged fasting.


Time of year

Early January seemed like a good time for a few reasons. First, no one eats right or exercises consistently around the holidays, so we usually have a few pounds to lose.

Second, January is at the tail end of “off season” for triathlon, so experimenting with fasting then is less disruptive to training.

Finally, it’s easy to start a nutritional plan at the start of the year when everyone is talking about resolutions. Even easier when it only lasts two weeks.

One reason NOT to do the fast in January: temperature. While fasting, BMR drops as your body conserves resources to keep you alive as long as possible. If you live in a cold climate, you may find the fast less unpleasant in the summer.

Time of week

We planned to fast during the week so we’d have plenty to do: thinking about work means less thinking about food. We also wanted to avoid interrupting our usual weekend long bike rides and runs.

Time of day

It’s easy to fast while sleeping, so it seemed like a good “bang for the buck” to start the fast after dinner and break the fast with breakfast.


We sandwiched our fast between two periods of keto, a protocol Dr. Peter Attia refers to as the “Nothing burger”:

  1. One week, Monday to Sunday, of strict keto, with Sunday dinner being our last meal before the fast

  2. Three full days and four sleeps (84 hours) of water and minerals only

  3. Break the fast Thursday morning with strict keto throughout the rest of the week

On the last Sunday, we celebrated completing our program with all the carbs our hearts desired.

Why keto?

On the frontend, we wanted to enter ketosis before starting the fast, which we thought might make the transition more pleasant (or at least less unpleasant). I don’t have a reference point for comparison, but our friend Barry, who joined us for the fast and has prior fasting experience, says his hunger levels were much more tolerable this time around.

On the tail end, we wanted to avoid overeating, refeeding syndrome, and possibly ride out any additional benefits that may come with maintaining ketosis a few more days post-fast. There’s not good research on how best to break a fast, but it seems like good ideas include eating small portions, staying hydrated, and choosing nutritious foods.

You can see the kinds of included foods on our keto grocery list.

Fluids and electrolytes

To stay hydrated and avoid some commonly reported symptoms like cramping, fast heart rate, and insomnia, we regularly drank water with electrolytes. We usually get a fair bit of water from the food we eat, so while fasting, one must drink more fluids than usual.

I started every morning with a piping hot cup of water with a pinch of salt added. It sort of fulfilled the need for a morning ritual, albeit less satisfying. Throughout the day we kept water bottles handy with about 1/8 teaspoon of salt per 8 ounces of water.

Per the recommendation of Attia, we took 200 mg of magnesium (as magnesium glycinate) twice daily, once in the morning and once in the evening. Magnesium supposedly helps reduce cramping and aids sleep. Attia warns not to take magnesium oxide as it can cause GI distress.

Beyond water, salt, and magnesium, we did not consume anything else. We had potassium salt at the ready but elected not to use it as we did not experience cramping.


Yes, you can (and probably should) exercise during the fast. It’s counterintuitive, but exercise will help preserve your skeletal muscle instead of catabolizing it as much. The mechanism is beyond the scope of this article, but the gist is that exercise stimulates mTOR, and mTOR helps maintain skeletal muscle.

B and I chose to follow our normal workout routine but reduce the intensity quite a bit. During the fast, we biked Monday, jogged and swam Tuesday, and biked again Wednesday. Each day we also went for a walk, primarily during mealtimes. We also used stretch cords and body weight exercises to work the large muscle groups.

It’s challenging to stay hydrated and consume enough minerals while fasting, so it’s advisable to avoid workouts that result in heavy prolonged sweating. Strength sessions are probably a good idea, but be sure to increase time between sets.

Does X break the fast?

Some people allow black coffee or unsweetened tea during a fast. Others allow non-caloric electrolyte powders with artificial sweeteners. Though I don’t know of any evidence one way or the other, I’m erring on the side of caution and simulating a fast as close to what caveman-era man would have experienced. Pretty sure they didn’t have Americanos or Splenda back then.

That said, the first-order goal should be zero caloric consumption*. If the choice is fast with coffee or not at all, obviously the former is better. Since I thought I could finish the fast without caffeine, I opted to forgo my daily cup o’ Joe.

It could very well turn out that caffeine or other chemicals upregulate autophagy while fasting. I expect that as researchers discover more, I will update my attitudes on this matter.

*Since insulin and amino acids reduce autophagy, one should avoid carbohydrates and protein. Eating pure fat, however, may have little to no negative impact despite having calories. Non-nutritive sweeteners, despite having no calories, may affect autophagy. There could be mechanisms at work we don’t yet understand, so as I already mentioned, I’m attempting to fast in the same way our ancestors may have fasted.

Setting ourselves up for success

We did a few things to help ensure we’d make it through the fast:

  • Avoiding tempting food environments. It’s easier to fast when you’re not constantly reminded of food. Keep water and salt handy but put everything else away. If you live with others who aren’t fasting, go for a walk when they eat, and ask them politely to avoid cooking fragrant foods.

  • Have friends/family who support you. Besides having each other, B and I recruited our friend Barry to join us for the fast, providing valuable emotional support. If it’s not possible to fast with a friend, consider joining an online support group or using the Zero fasting app.

  • Have a backup plan if things go south. Bailing is always an option, but we included the following alternatives to help avoid complete failure:

    1. Permit caffeine

    2. Allow chicken broth

    3. Add some fat, e.g. heavy cream in coffee

    4. Allow up to 10% daily calories

    5. Return to keto

  • Plan social activities during mealtimes. We are creatures of habit, and having planned activities for mealtimes helps distract the inevitable feeling of “I should be eating right now”. I primarily went for walks and talked on the phone with friends and family during mealtimes.


Day 1

Morning readings (Zack and B):

  • Weight (lbs): 174.4 and 112.6

  • Glucose (mg/dL): 83 and 78

  • Ketones (mmol/L): 0.7 and 2.2

The morning felt like a normal morning. On occasion I have only coffee for breakfast, so drinking hot water felt normal.

In the early afternoon, I started to notice feeling “empty”. It didn’t feel like hunger per se, but I felt like I was missing something. I went for a walk around the block, which helped a lot.

Later in the afternoon I felt the emptiness again. Whenever the feeling became sufficiently distracting, I did some air squats, pushups, or situps, as if movement is the off-switch to hunger signals. It (mostly) worked.

Around dinner time, the emptiness feeling was more noticeable. I hopped on the bike and Zwifted for about an hour at a chill pace. I added a couple burst efforts to see how they would go - lower power than normal, but not bad.

By bedtime, I felt great. Zero hunger or discomfort of any kind. I closed my eyes and fell asleep quickly.

Day 2

Morning readings (Zack and B):

  • Weight (lbs): 172.2 and 112.0

  • Glucose (mg/dL): 89 and 72

  • Ketones (mmol/L): 2.0 and 3.9

We started our day intending to do a relaxed run. We started with an uphill portion, immediately resetting our expectations on how the run was going to go. Anything over a 1% grade resulted in near instant zone-5 heart rate.

Realizing a steady state run was out of the picture, we opted for run/walk intervals, 1 minute on, 1 minute off. Sad, but it’s what we could manage. We did a short three-mile loop and called it quits.

Everything cognitively seemed typical while working. I noticed a little bit of hunger once or twice, but it passed quickly. When standing up, I almost always had tunnel vision, a good reminder to drink some salted water.

In the afternoon, we did a swim workout with the UCSD tri team. The workout was centered around a 500 yard time trial. We told our coach in advance what we were doing and that we planned to take the TT easy.

I swam the 500 at a pace I knew I could finish comfortably, but not slow either. I checked the clock after the first 200. 2:25. “Wow, really?” I thought to myself. I was surprised I could hold a decent pace, especially after such a weak run earlier in the day.

I came into the wall at 5:58, shocked that my body, depleted of glycogen, could manage a sub-6 minute 500.

Home from swim, both of us were noticeably colder than normal. We cranked the heat, added layers, and sipped hot water. We spent the rest of the evening curled up under blankets.

Day 3

Morning readings (Zack and B):

  • Weight (lbs): 172.0 and 111.0

  • Glucose (mg/dL): 86 and 57

  • Ketones (mmol/L): 2.4 and 6.1

I woke up an hour before my alarm feeling very awake. Strangely, I slept less than normal, but I still felt rested. B woke up with pain in her lower back. She described it as coming from her tailbone and deeper than her muscles. We took numbers and headed out for a short breakfast walk. It was lovely sunny San Diego morning.

Health bloggers often report improved cognition or even euphoria while fasting. Three days in, neither of us experienced any cognitive changes for the better. B described her mental state as “high without the dopamine.” I found I could focus on mentally demanding tasks, like reading research papers, for less time than usual.

We were not hungry, but we could feel our empty stomachs, which made noises on occasion. Time seemed to move slower than usual, like on a long flight. We knew it would end eventually but the end was not coming fast enough.

I’m not sure which forces, physiological or psychological, were dominant at this point in the fast. Either way, we were both looking forward to having food again. B was craving some form of fried potatoes and I was craving pastries and Chinese food. Or just about any food, really.

We did another Zwift workout, targeting a zone-2 pace. We managed to produce about 50-75 fewer watts versus baseline for the same heart rate. Still, I was pleasantly surprised that after not eating for 72 hours we could still manage 75% our typical load.

As bedtime rolled around, I was not feeling sleepy. I tried the normal tricks - reading, meditating, breathing exercises - but nothing worked. I gave up and returned to work until I started yawning a few hours later.

Day 4 - breaking the fast

Morning readings (Zack and B):

  • Weight (lbs): 168.4 and 109.0

  • Glucose (mg/dL): 81 and 52

  • Ketones (mmol/L): 3.1 and 6.2

At last, it was time to break the fast! Like impatient children on Christmas morning, we woke up a little earlier than usual and rushed downstairs.

First things first: we took final weight and blood measurements. My ketone trend continued its slow creep upwards while B maintained her already high levels.

Next: coffee. I made my usual Americano, took a sip. “B, this is the best cup of coffee I’ve ever had.” Not kidding.

We made our usual keto breakfast of two eggs, spinach, tomato, and avocado. I added some lox to my plate. First bite…a symphony of flavors. I’ve made this breakfast hundreds of times, but something was different. Of course, it wasn’t the food that was different. It was us. Regardless, I enjoyed every bite, trying to savor it despite my instincts to gobble it down.

With my belly full, I noticed B has barely touched her food. “You okay?” She was feeling nauseous. We warmed up some bone broth, which seemed to help. After a while she was able to stomach the rest of her breakfast, but she didn’t want to keep eating. I, on the other hand, knew I could eat more despite feeling very full. I forced myself to wait a couple hours, hoping to escape GI issues.

Throughout the day we ate smaller, frequent meals, all of which tasted better than usual, a rather wonderful unexpected side effect.

We rode our bikes in the afternoon and reported back feeling hella good. Power numbers were still lower than baseline, but not low for keto. Climbing Torrey Pines, I felt lighter than usual, but it could have been the placebo effect .

Cheat day

Now for some fun. This section is just about our food, though B did take blood measurements to confirm she finished the week in ketosis.

  • Breakfast: bagels

  • 2nd breakfast: donuts

  • Lunch: The Habit burgers, fries, and tempura green beans

  • Snack: Croissants from Arely’s Bakery

  • 2nd snack: Slice of Julian Pie (thanks Rory!)

  • 3rd snack: Chocolate chip cookies (thanks Barry!)

  • Dinner: a little bit of rice and chicken

  • Dessert: Mochi

  • Throughout the day: lots of fluids, a couple scoops of Metamucil (essential on cheat day!)

I’m wrapping up writing this (on cheat day) feeling satisfied and about to explode 😋

Parting thoughts

Fasting gets a lot of hype. The scientific community has provided some evidence that it’s “probably” beneficial for human health, but there are no vetted guidelines on how, how long, or how often one should fast to prevent disease. It’ll probably be too late for us by the time we have good guidelines, so I’m being bullish now given that the potential risks are small and potential upside is large.

This was B’s first time fasting and her first time following a ketogenic diet. Interestingly, her ketone numbers were consistently elevated during the fast. I am curious to know if her high levels could be explained by fast rates of ketone production or, perhaps, slow rates of metabolism. Blood meters tell us levels but not flows of ketones.

It’s possible that with some more time to adapt, B’s fasting experience might have been even less unpleasant. I know she’s not eager to fast again, but if she does, it’ll be interesting to see if her cellular machinery is now better prepared for ketosis.

For me, the experience wasn’t a cake walk, but it wasn’t terrible either. I didn’t experience a lot of what people claim like exceptional mental clarity, feeling connected to the universe, or faster healing of aches and pains. Hopefully I upregulated autophagy.

I wouldn’t want to do it every month, but once a year seems manageable. Until then, I’ll be enjoying some scrumptious not-fasting.

This post contains affiliate links, which means I may earn a small commission if you click on one and make a purchase, at no extra cost to you. Thank you for supporting the site!