BLUF:
I tested the Fasting Mimicking Diet - 1 day at 1090 kcal, then 4 days at 720 kcal with a macro split that's roughly 45% carbohydrates, 45% fats, and 10% protein - and report back that it may have been helpful in achieving body recomposition (reducing body fat percentage; increasing lean mass), and was a tolerable experience.Background:
Numerous studies have touted the medical utility of of every other day fasting (e.g. 24 hours of 0 kcal, alternating with 24 hours of ad libitum eating) or similar diets against a range of clinical outcomes from weight loss to autoimmune disease [1]. But not eating anything for 24 hours seems very challenging. The Fasting Mimicking Diet (FMD) is designed to achieve similar physiological results, but with a regimen that's hypothetically easier to adhere to: 1 day at 1090 kcal, then 4 days at 720 kcal with a macro split that's roughly 45% carbohydrates, 45% fats, and 10% protein.
The first major FMD paper came out in 2015 with a range of impressive outcomes in mice who underwent several cycles of FMD: durable weight loss (i.e., weight that stayed off after they stopped FMD cycles), longevity (11% longer lifespan), enhancement in cognitive performance, immune system rejuvenation (e.g. lymphoid/myeloid progenitor ratio up). The weight loss appears to mainly target adipose tissue while leaving lean body mass intact. It also looks like each cycle massively shrinks your organs - liver, heart, kidneys - which return to normal mass on refeeding [2]. In another study, FMD regenerates pancreatic β cells and cures T1D [3], and in yet another it slows down MS onset and reduces disease severity (and there's some broader speculation about autoimmune disease in general) [4]. But as Gary Taubes once mentioned to me, "if you can't cure cancer in mice, you probably shouldn't be in this business". So what about humans?
There's a few human trials. The biggest study with published results is from 2017 and in Science Translational Medicine. Compared to a control, after 3 cycles of FMD (3x 5-day fasts, each with about 25 days of normal eating in between), FMD participants (measurements taken 5-7 days after the last fasting cycle) achieved substantial weight loss, lower systolic BP, and serum IGF-1, which is a hormone associated with aging and cancer. Remarkably it looks like the same body composition effect we can see in mice - growing proportion of mass that is lean tissue - also occurs in humans, although we are really at the edge of detection [5].
In the MS study there was a small human arm where a single cycle of FMD followed by Mediterranean diet for 6 months improved quality of life scores over control and in some cases ketogenic diet patients with relapsing-remitting MS (RRMS), though no difference in relapse rate [4]. The authors also observed a substantial (20%) drop in lymphocyte count [4]. There is also some early thinking around protection from chemotherapy sensitization in humans [6].
The first major FMD paper came out in 2015 with a range of impressive outcomes in mice who underwent several cycles of FMD: durable weight loss (i.e., weight that stayed off after they stopped FMD cycles), longevity (11% longer lifespan), enhancement in cognitive performance, immune system rejuvenation (e.g. lymphoid/myeloid progenitor ratio up). The weight loss appears to mainly target adipose tissue while leaving lean body mass intact. It also looks like each cycle massively shrinks your organs - liver, heart, kidneys - which return to normal mass on refeeding [2]. In another study, FMD regenerates pancreatic β cells and cures T1D [3], and in yet another it slows down MS onset and reduces disease severity (and there's some broader speculation about autoimmune disease in general) [4]. But as Gary Taubes once mentioned to me, "if you can't cure cancer in mice, you probably shouldn't be in this business". So what about humans?
 |
| Adapted from Brandhorst (2015): FMD cycles in mice result in weight loss, particularly targeting adipose tissue. SAT is subcutaneous and VAT is visceral. Lean body mass stays intact. |
There's a few human trials. The biggest study with published results is from 2017 and in Science Translational Medicine. Compared to a control, after 3 cycles of FMD (3x 5-day fasts, each with about 25 days of normal eating in between), FMD participants (measurements taken 5-7 days after the last fasting cycle) achieved substantial weight loss, lower systolic BP, and serum IGF-1, which is a hormone associated with aging and cancer. Remarkably it looks like the same body composition effect we can see in mice - growing proportion of mass that is lean tissue - also occurs in humans, although we are really at the edge of detection [5].
 |
| Adapted from Wei (2017): FMD cycles in humans result in weight loss, and like mice it appears that there is some kind of body recomposition effect here. I'm guessing there was some p-hacking going on here (delta Rel volume is a weird y-axis) but it looks like there's something to it. |
In the MS study there was a small human arm where a single cycle of FMD followed by Mediterranean diet for 6 months improved quality of life scores over control and in some cases ketogenic diet patients with relapsing-remitting MS (RRMS), though no difference in relapse rate [4]. The authors also observed a substantial (20%) drop in lymphocyte count [4]. There is also some early thinking around protection from chemotherapy sensitization in humans [6].
Experiment:
Objectives
It seems like FMD may hold therapeutic potential in humans but obviously we are not yet at a place where we have good evidence - this will only come with more well-designed trials. There does seem to be a more robust, although still unproven, relationship between FMD and body recomposition.
My objectives here were therefore twofold:
- (1) Body mass recomposition. Can I achieve a similar weight to baseline on refeeding, but higher ratio of lean mass to adipose tissue?
- (2) Test the tolerability of FMD. If this is something that I think most humans could tolerate? Sick humans?
Intervention
The "real" FMD is available for purchase from a company that the scientists behind FMD started. I hacked together something that got the calorie count and macro ratios roughly right with some help from a fellow experimenter's records.
Actual daily intake records are:
|
|
Calories
|
Carbohydrates
|
Fat
|
Protein
|
|
Day 1
|
1120
|
45.2%
|
43.7%
|
11.1%
|
|
Day 2
|
715
|
47.1%
|
42.2%
|
10.7%
|
|
Day 3
|
715
|
47.1%
|
42.2%
|
10.7%
|
|
Day 4
|
715
|
47.1%
|
42.2%
|
10.7%
|
|
Day 5
|
715
|
47.1%
|
42.2%
|
10.7%
|
|
8:00 AM
|
Coffee with Cinnamon (10 oz)
|
0
|
0
|
0
|
0
|
|
8:50 AM
|
Whole Almonds (2 tbsp)
|
85
|
3
|
7.5
|
7
|
|
8:50 AM
|
Lundberg Brown Rice Cakes (1 cake)
|
60
|
14
|
0.5
|
1
|
|
8:50 AM
|
Fish Oil Pill
|
10
|
0
|
1
|
0
|
|
8:50 AM
|
Green Superfood (8 g / 1 scoop)
|
30
|
4
|
0
|
2
|
|
11:00 AM
|
Lundberg Brown Rice Cakes (1 cake)
|
60
|
14
|
0.5
|
1
|
|
11:00 AM
|
1/4 tsp Sea Salt in water
|
0
|
0
|
0
|
0
|
|
12:00 PM
|
Full Hass Avocado
|
250
|
13
|
23
|
3
|
|
7:45 PM
|
Trader Joe’s Chicken Broth (2 cups) with Cumin sprinkled on top
|
40
|
4
|
2
|
0
|
|
7:45 PM
|
Lundberg Brown Rice Cakes (1 cake)
|
60
|
14
|
0.5
|
1
|
|
7:45 PM
|
Diced Tomatoes (14.5 oz / 1 can)
|
90
|
18
|
0
|
3
|
|
7:45 PM
|
Green Superfood (8 g / 1 scoop)
|
30
|
4
|
0
|
2
|
|
8:15 PM
|
Chamomile Tea with Cinnamon
|
0
|
0
|
0
|
0
|
Or visually:
 |
| A typical day, less a snack rice cake and my ad hoc chamomile teas. The avocado would typically take some weighing and trimming to get to 5 oz. |
You are supposed to eliminate coffee for the FMD but give my high typical intake (about 4 10 oz cups a day) I thought that would not be prudent. I limited myself to one strong cup brewed each morning that I would keep in a thermos and sip on through the early afternoon.
During the FMD I also abstained from all exercise and rigorous physical activity.
After the fast ended I reverted back to my typical diet - which emphasizes whole plant foods, grass-fed meats, sardines, and flavored, carbonated water, and excludes liquid calories (except occasional alcohol), added sugars, artificial sweeteners, desert foods, trans fats, and anything that is burnt. I ate ad libitum. I do not have great records but I would peg my daily intake at about 2600 kcal/day which is slightly higher than my normal maintenance intake of about 2400-2500 kcal/day. At no point did a really have the urge to gorge myself, so I didn't. I also returned to normal patterns of physical activity, lifting 3 times / week and running about 10 miles / week.
Results:
(Outcome 1) Body Mass and Composition
Readings taken with 10 minutes of waking each morning, after showering, on my Nokia Body Cardio scale. I trust the weight and generally do not trust the body fat % number - DEXA would be much more reliable.
At seven days of refeeding I am at a slightly lower weight than baselines. It also seems like by body fat percentage is slightly lower (concordant with my own subjective feelings from visual inspection). But I don't think that we can say anything conclusive here yet given the day to day variance and potential inaccuracy.
(Outcome 2) Tolerability
I recorded "feeling hungry" once in my log during day 2 of the fasting cycle. I also recorded nearly every day a feeling of coldness/clamminess. But no serious adverse events. Mental clarity was strong throughout - no superhuman focus, but I felt sharp! By the last day I actually recorded "this was much easier than I thought it would be". I think that FMD is tolerable in healthy humans.
Discussion & Unsubstantiated Claims:
I want to keep pushing on better data for body recomposition. I will do this by repeating cycles myself, paired with better measurement (working on DEXA scans). I would also love for others to test the diet and record results.I am fascinated by how quickly I lost and gained weight going on and off of the diet. If you look at caloric deficit during the fast and surplus after the fast, I am changing weight far more rapidly than the rule of thumb (3,500 kcal/lb) would have us think. Now maybe there is something going on with water weight but I would be suspicious if that explained it all.
Another piece is the pattern of weight loss. I don't think I lost much subcutaneous fat at all, but my abdomen and waist really shrunk down while fasting. This is pretty consistent with the studies in mice and humans [1, 5]. Seems like there is a real difference in how the body decides to make visceral fat vs. subcutaneous fat.
More broadly the results we are seeing around FMD and similar studies with ketogenic diet, IF, etc should make us intrigued about the potential of nutritional interventions for obesity and diabetes, autoimmune disease, and oncology. As an example of practical implementation, Geisinger has achieved very strong health outcomes and substantial PMPM savings for diabetic members by "prescribing" food deliveries for members and families. Gottlieb's FDA has similarly made some statements around creating frameworks to enable the makers of food products to make stronger health claims, in an attempt to spur innovation and the development and marketing of more healthful foods - this actually appears as item 3 on their strategic plan for 2018.
References
[1] Mattson et al., PNAS (2014)[2] Brandhorst et al., Cell Metab. (2015)
[3] Cheng et al., Cell (2017)
[4] Choi et al., Cell Reports (2016)
[5] Wei/Brandhorst et al., Sci Transl Med. (2017)
[6] Di Biase et al., PLoS Biol. (2017)
