One of the most common concerns about a long fast is losing muscle mass. Yes, we do lose muscle mass, but few people know how much they actually lose. The usual answer is a scale number or a Dexa lean mass number. These numbers don’t answer the muscle protein question. This post is about muscle protein - the stuff that actually contracts - the most critical part of lean mass. So let’s talk about muscle protein loss during extended fasts. I will show you my own Dexa scans from two 10-day fasts and give you my estimate of what one fast costs and whether you get it back.
📖 Three Terms, Three Different Things
Let’s start with terminology. Different terms are used when people talk about muscle loss, and they don’t mean the same thing.
Lean mass - total weight minus fat minus bone, which is how Dexa providers like BodySpec define it. Muscle, but also organs, skin, blood, gut contents and all the water in your body. This is what a Dexa scan or smart scale reports. During a fast most of the change in lean mass is water and glycogen, not muscle.
Muscle mass - the wet weight of skeletal muscle. About 20% protein, 75% water, and a few percent glycogen, fat and minerals. Muscle mass can drop a lot while muscle protein barely moves, because water and glycogen leave first.
Muscle protein - the actual contractile machinery inside muscle fibers, mainly actin and myosin, plus the enzymes and mitochondria that run them. This is what does the work, and this is what the post is about. Roughly 20% of a muscle’s weight.
So the chain goes like this. A Dexa shows lean mass, lean mass contains muscle mass, and muscle mass contains muscle protein. Each step down is a smaller number, and each step up adds water.
🔬 Why Muscle Protein Loss Is Hard to Measure
One of the complications is that muscle protein loss cannot be measured at home. Your scale does not see it. A Dexa scan does not see it either - it sees lean tissue, which is water, glycogen, organs and muscle lumped together.
The real tools live in research labs:
24-hour urinary nitrogen - the oldest method and the one behind most numbers in this post. Protein is the only fuel that contains nitrogen, so nitrogen in the urine tracks protein burned. 1 g of nitrogen = 6.25 g of protein. A small amount also leaves in sweat, skin and hair, so urine slightly understates the true loss, especially on heavy sweat days.
3-methylhistidine - an amino acid released only when contractile muscle protein is broken down. It is the closest thing to a muscle-specific breakdown marker.
Amino acid tracers - labeled phenylalanine infused into the blood, with catheters in an artery and a vein of the forearm, to measure how much protein muscle releases and takes up.
Muscle biopsies - a needle sample of thigh muscle to look at signaling and fiber structure.
MRI and MR spectroscopy - muscle volume, with corrections for glycogen and water.
None of this is practical outside a study. So every number below comes from a handful of trials, most of them small. Keep that in mind - these are the best numbers we have, not the final word.
📜 What Cahill Established
The modern understanding of fasting metabolism comes from George Cahill at Harvard. His 1970 review in the New England Journal of Medicine, “Starvation in man”, is still the reference everyone cites. He built on Francis Benedict’s 1915 study of a man who fasted 31 days, and on his own group’s work with fasting volunteers in the late 1960s.
Cahill’s framing goes like this. Nitrogen is never stored. Unlike fat in a fat cell or glycogen in the liver, every molecule of protein in your body is doing a job - an enzyme, a structural fiber, a contractile unit. So every gram of protein lost during a fast is working machinery given up.
Here is the problem the body has to solve. Your brain burns about 140 g of glucose a day. Fat cannot be turned into glucose. So in the first days of a fast, the glucose has to come from amino acids, and most of those come from muscle. Cahill’s numbers for a normal man after 1 to 2 days without food - about 75 g of protein a day mobilized, mainly from muscle, plus 160 g of fat. That is a clearly negative nitrogen balance.
Then the body adapts. The liver turns fat into ketones and the brain switches to ketones for two thirds of its fuel. Its glucose demand drops to about 40 g a day. Add the red blood cells and a few other tissues that can only run on glucose, and the whole body gets by on about 80 g a day (Owen 1969). Most of that comes from recycled lactate and from glycerol released by fat. Only the rest has to come from protein. Urinary urea, the main form of lost nitrogen, falls to 1 to 2 g a day. Without this switch, the body would keep burning about 75 g of protein a day, and a few weeks of fasting would cost several kilograms of muscle. The switch to ketones cuts protein loss to roughly a third of that, and this is what makes long fasts survivable.
His summary - the body’s maneuvers to spare nitrogen are so efficient that fat becomes the main fuel and muscle mass is largely maintained.
And when food comes back, the balance flips. Refeeding after a long fast produces a positive nitrogen balance - the body retains more nitrogen than it excretes and rebuilds what it used (Marliss 1978, refeeding with a protein diet after 21 to 28 day fasts).
🧬 Why Muscle Releases Protein
Muscle supplies the two amino acids the liver wants most for making glucose - alanine and glutamine. In the first days, muscle releases them at a high rate.
What happens inside the muscle is more specific than “muscle breaks down”. A 2014 study put 8 healthy men through a 72-hour fast with forearm tracers and biopsies (Vendelbo 2014). Net protein release from muscle went up. But the main breakdown enzymes did not change. What changed was the building side - mTOR signaling, the master switch for protein synthesis, dropped by about 50%. Muscle loses protein during a fast mostly because it stops building, not because it starts tearing itself down faster.
Why does the loss slow down after a few days? The classic answer is ketones. In 1975, Sherwin infused ketones into people who had fasted 5 to 10 weeks and urinary nitrogen fell by 30% (Sherwin 1975). The counterpoint came in 1980 - Féry and Balasse showed that the sodium salt used in those infusions raised blood pH, and that explained the effect, while ketones in their natural acid form did the opposite (Féry 1980). We know protein sparing happens as ketosis deepens, but the exact signal is still debated 50 years later.
📉 How Much Muscle Protein Is Lost
Now the numbers. These are estimates built from the nitrogen studies, and I will show the arithmetic.
Total protein first. The best day-by-day data comes from a 1975 study of 72 people - 24 of normal weight fasting 6 days and 48 obese fasting 6 to 28 days (Goschke 1975). Urinary nitrogen rose from day 1 to day 3, then fell steadily. Daily losses ranged from 14.5 g of nitrogen (men, early in the fast) to 3 g (obese women, in week 4). Protein’s share of total energy fell from about 15% in a 6-day fast to 5% by week 4.
For a 10-day fast that works out to roughly 75 g of protein a day in the first days, falling to about 30 g a day by day 10. Total - around 450 g of protein.
Muscle’s share. Not all of that comes from muscle. The liver, gut lining and kidneys turn over protein much faster than muscle and give up a large share early. Muscle holds about half of the body’s protein and supplies most of the early alanine and glutamine. A reasonable estimate is that muscle contributes 40 to 50% of the total.
That gives about 200 g of muscle protein over a 10-day fast.
Muscle is roughly 20% protein and 75% water, so 200 g of protein is about 0.8 to 1 kg of wet muscle, or 2 lbs. An average man carries about 30 kg of skeletal muscle. So a 10-day fast costs about 3% of your muscle.
By day it looks like this - about 30 g of muscle protein a day on days 1 to 3, then falling, with the turn around day 5. That turn shows up directly in the data. In a 2021 trial of 16 men fasting 10 days, plasma 3-methylhistidine, the muscle breakdown marker, rose until day 5 and then fell (Laurens 2021).
Two caveats. First, no study has measured muscle protein directly over a 10-day fast, so 200 g is an informed estimate, not a measurement. Second, these numbers come from people fasting at rest or with light activity. Hard training during a fast changes the picture.
🏋️ Muscle Function Is Preserved
The muscle lost on paper does not show up in what the muscles can do. In the 10-day trial, strength in the arms, the non-weight-bearing muscles, was unchanged, and strength in the legs went up by 33% - the men walked up to 3 hours a day during the fast (Laurens 2021).
A 2025 study went further. 32 people, half of them over 50, fasted 12 days on 250 kcal a day while researchers tracked their muscles with MRI and spectroscopy (Naëgel 2025). Calf muscle volume dropped 5.4%, but that drop closely matched the expected loss of glycogen and the water bound to it. Mitochondrial function was unchanged. Maximal strength was preserved in both thighs and calves, regardless of age or sex. Fat oxidation during exercise improved. Women’s mitochondrial recovery improved by day 12.
Their conclusion - human muscle, like animal muscle, seems to have evolved to withstand long periods without food.
⚖️ Lean Mass - A Proxy That Overstates the Loss
Get a Dexa scan before and after a long fast and you will see 3 to 4 kg of lean mass gone. That number is real, but most of it is not muscle.
In the 10-day trial, lean soft tissue dropped 3.53 kg and made up 60% of total weight loss (Laurens 2021). The researchers broke that 3.53 kg down - 44% extracellular water, 14% glycogen with its bound water, and 42% metabolically active tissue. So about 1.5 kg of real tissue, spread across liver, gut, kidneys, heart and muscle. Convert that to protein and it is 300 to 400 g, which lines up with the nitrogen math above. Muscle is a fraction of it.
A 2025 water-fasting study with Dexa found fat-free mass was 74% of the weight lost by the end of a 14-day fast (Gabriel 2025). A 2026 meta-analysis of water-only fasting trials found the same pattern across studies and the same explanation - fat-free mass estimates are driven by hydration, and much of the loss returns during refeeding (Nutrition Reviews 2026).
So the general statement is this. Lean mass is a proxy, and a poor one during a fast. It tracks water and glycogen first, organ tissue second, and muscle protein last.
📊 What My Dexa Shows
I take a Dexa scan before each fast, right after, and again after refeeding. Here are my two 10-day fasts, in pounds.
10-Day Water Fast, Aug-Sep 2025
10-Day Water Fast, Jul-Aug 2026
Both fasts show the same pattern. Lean tissue dropped about 8 lbs (3.6 kg) during the fast - almost exactly the 3.53 kg from the 10-day trial. Then 12 days of refeeding brought it back. Net lean change after one full cycle - minus 0.3 lbs in 2025, plus 1.3 lbs in 2026. Fat dropped 4.2 and 4.5 lbs net and stayed off.
If the 200 g estimate is right, I lost about 2 lbs of actual muscle during each fast, hidden inside the 8 lbs of lean loss. By the refeed scan it was back, within the noise of a Dexa scan. My full fast data is on the results page.
🍽️ Recovery During Refeeding
The nitrogen balance arc is the answer. Negative during the fast, positive during refeeding. The water and glycogen come back in days. In the 10-day trial, myostatin, the hormone that brakes muscle growth, dropped significantly during food reintroduction, which clears the way for rebuilding (Laurens 2021). At the 3-month follow-up, body weight was still 2.3% below baseline and lean tissue 3.2% below, so the lean fraction was about the same.
The 2025 Dexa study followed people for 6 weeks after a 14-day water fast (Gabriel 2025). At the end of the fast, fat-free mass was 74% of weight lost. At follow-up, total weight was down 8.2%, fat-free mass down only 4.4%, and fat mass down 14.7%. Lean recovered while fat kept falling. And that was in people with a median age of 62.
For a 10-day fast, 200 g of muscle protein is what a normal diet with enough protein, plus a return to training, rebuilds in 2 to 4 weeks. My Dexa scans say 12 days was enough.
🧮 The Net Effect of One Fasting and Refeeding Cycle
Here is my estimate for one fasting and refeeding cycle.
Muscle protein - neutral to slightly negative. You lose about 200 g during a 10-day fast. Most of it comes back with a protein-forward refeed and training. Some people end slightly below where they started, some end even. No study has shown muscle protein going above baseline after a fast in humans - claims that you “come back stronger” are extrapolated from animals and growth hormone spikes, not measured.
Is that loss harmful? Not necessarily. Fasting upregulates autophagy and, in other tissues, apoptosis. For muscle the relevant process is autophagy - clearing damaged mitochondria and misfolded proteins inside the fiber. Autophagy markers in human muscle rise within 72 hours of fasting (Vendelbo 2014). Muscle fibers do not die and get replaced like gut cells, so apoptosis is not the main story here.
The hopeful reading is that some of the protein you lose is the damaged kind, and what you rebuild is cleaner. Whether that is true has never been measured in humans. Plausible, unproven.
What is measured - function holds or improves, mitochondria stay intact, and fat oxidation gets better. So the trade looks like this. A small, recoverable cost in muscle quantity, a possible gain in muscle quality, and a real gain in metabolic flexibility.
That is a reasonable trade if you have solid muscle reserve and you refeed with protein and train. It is a poor trade if you are older with low muscle mass, where quantity itself protects you from falls and frailty.
🛡️ My Muscle-Protection Strategy
What I do around a 10-day fast to keep the cost small.
Gym almost every day. Working out early drains glycogen and gets me into ketosis faster, which shortens the high-protein-loss window. The intensity of my workouts goes down quite a lot after day 3.
Be physically active. The 10-day trial men increased their steps by 60% and their leg strength went up. Load on weight-bearing muscle is a signal to keep it.
Electrolytes from day 1. Potassium and sodium loss tracks nitrogen loss at a fixed ratio. My protocol is in Electrolytes for Fasting.
Cap at 10 days. Protein loss per day keeps falling, but so does the marginal benefit, and risks rise past 7 to 10 days. The fasting timeline shows what changes when.
Protein-forward refeed. Broth and vegetables first, then protein at every meal by day 3. The full sequence is in How to Break an Extended Fast.
Start ramping up my workout load after breaking the fast. This is when the positive nitrogen balance does its work.
If you want a projection of your own lean versus fat loss before a fast, the fasting weight loss calculator estimates both.
⚠️ Who Should Be Careful
The studies above were done in healthy, mostly middle-aged men and women with normal muscle mass. The numbers do not transfer to everyone.
Older adults with low muscle mass. 3% of a small reserve matters more, and rebuilding is slower. The Buchinger Wilhelmi researchers say it directly - for older underweight people with muscle loss, other strategies are indicated.
Very lean people. Less fat means the body shifts back toward protein for fuel sooner.
Back-to-back fasts. Doing another long fast before the rebuild is done stacks the losses.
Medication, diabetes, pregnancy, eating disorders. Doctor first, not a blog post.
✅ Final Takeaway
One 10-day fast costs about 200 g of muscle protein, roughly 3% of what you carry. Your Dexa will show 8 lbs of lean loss, and most of that is water and glycogen that returns within days. Strength and mitochondria hold. Refeeding flips nitrogen balance positive, and with protein and training the muscle comes back in weeks.
One fast costs a little muscle protein. Refeeding and training decide whether you get it back.
More on what fasting does to the rest of the body in Fasting and Autophagy and my full 10-Day Water Fast report.
📚 References
Muscle protein loss during fasting has been studied directly in only a handful of human trials, but the physiology has been understood since the 1960s. Here are the papers that back up the key numbers in this post.
“Starvation in Man” 🔗 - Cahill’s classic review of how the body shifts from glucose to ketones during fasting and spares protein in the process.
“Liver and Kidney Metabolism During Prolonged Starvation” 🔗 - Shows glucose production falling to about 86 g a day after 5 to 6 weeks of fasting, split between liver and kidney.
“Nitrogen Loss in Normal and Obese Subjects During Total Fast” 🔗 - Day-by-day nitrogen loss in 72 people, peaking on days 1 to 3 and falling steadily, the source for the protein loss arithmetic.
“Fasting Increases Human Skeletal Muscle Net Phenylalanine Release and This Is Associated with Decreased mTOR Signaling” 🔗 - Shows that muscle loses protein during a 72-hour fast mainly because protein synthesis slows, not because breakdown speeds up.
“Is Muscle and Protein Loss Relevant in Long-Term Fasting in Healthy Men?” 🔗 - The 10-day trial behind the lean mass breakdown, the day 5 turn in muscle breakdown markers, the strength results and the myostatin drop on refeeding.
“Impact of Long-Term Fasting on Skeletal Muscle: Structure, Energy Metabolism and Function Using 31P/1H MRS and MRI” 🔗 - MRI study showing that muscle volume loss during a 12-day fast matches glycogen and water loss, with strength and mitochondria preserved.
“Prolonged Water-Only Fasting Followed by a Whole-Plant-Food Diet Promotes Fat-Free Mass Recovery and Continued Fat Mass Loss” 🔗 - Dexa follow-up showing fat-free mass recovering after a 14-day water fast while fat loss continued.
“Duration-Dependent Changes in Body Composition During Prolonged Fasting: A Systematic Review and Meta-Analysis” 🔗 - Pooled analysis confirming that fat-free mass loss during water fasting is driven by hydration and partly recovers during refeeding.