Glycogen depletion and gluconeogenesis

Glycogen is glucose stored as a branched polymer, held hydrated with water, which is why the first pounds lost on any carbohydrate-restricted regimen come off fast and are mostly water rather than fat.

There are two separate depots and they are not interchangeable. The liver's glycogen can leave the liver and enter the blood. Muscle glycogen cannot leave the muscle, ever. That is not a matter of degree, it is a single missing enzyme, and it is explained below.

Gluconeogenesis is the manufacture of new glucose from things that are not glucose: the glycerol backbone released from fat, the amino acid alanine carried out of muscle, glutamine used especially by the kidney, and lactate returning from muscle. Only some of those represent genuinely new glucose carbon.

Rounded figures for how many grams each depot holds circulate everywhere. The research behind this page could not trace those gram figures to a primary source, so none of them are published here.

Breaking down glycogen in any tissue produces glucose-1-phosphate, which becomes glucose-6-phosphate. That molecule carries a charge, and the glucose transporters in a cell membrane carry only free, uncharged glucose. To export glucose a cell must first strip the phosphate off, and that requires the enzyme glucose-6-phosphatase.

Glucose-6-phosphatase is expressed in liver, kidney and intestine. It is not expressed in skeletal muscle. So a person can be carrying several hundred grams of glycogen in their legs and still become hypoglycaemic, because none of it can reach the blood. Muscle glycogen serves the muscle that stores it and nothing else. This single enzymatic fact is why liver and muscle stores should never be added together as carbohydrate available to the body.

Muscle glycogen does contribute indirectly, through the Cori cycle. Muscle runs glycolysis, pyruvate is reduced to lactate, lactate travels to the liver, the liver converts it back to pyruvate and runs gluconeogenesis, and glucose returns to the muscle. The accounting is unfavourable: glycolysis in muscle nets 2 ATP per glucose while hepatic gluconeogenesis costs 6 high-energy phosphate bonds. The cycle consumes net energy and shifts the burden from muscle to liver. Its purpose is not efficiency. And crucially, it is carbon neutral: the carbon arriving as lactate came from glucose in the first place. It recycles glucose carbon; it does not create any.

For genuinely new glucose carbon, four sources matter. Glycerol, released one molecule per triacylglycerol hydrolysed, is phosphorylated in the liver and enters the pathway directly. Alanine carries both carbon and nitrogen out of muscle in the glucose-alanine cycle. Glutamine is especially important as a substrate for kidney gluconeogenesis and simultaneously as the nitrogen carrier for renal ammonia production, which is how the kidney buffers the acid load of ketosis. Odd chain fatty acids yield propionyl-CoA and can produce net glucose, but they are a small minority of stored fat.

The glycerol point answers a question people ask constantly. The three fatty acid tails of a triglyceride cannot become net glucose in a human, because acetyl-CoA cannot be converted back to pyruvate and the two carbons that enter the citric acid cycle are lost as carbon dioxide before oxaloacetate is regenerated. The glycerol backbone can, and does, in quantity. In a 21 day fast, glycerol's contribution to gluconeogenesis equalled that of all amino acids combined.

Falling insulin and rising glucagon, acting on the liver through cyclic AMP and protein kinase A: glycogen phosphorylase is activated, glycogen synthase is inhibited, and the gluconeogenic enzymes PEPCK and glucose-6-phosphatase are induced. Cortisol is permissive, maintaining transcription of those enzymes and supplying amino acid substrate from muscle.

Almost every fasting timeline in circulation is built on the premise that glycogen runs out at some hour, and that this depletion is the switch that turns on fat burning, ketosis and everything downstream of it. The human measurement says gluconeogenesis is already supplying about two thirds of glucose production within the first day, confirmed by a second, unrelated method at roughly half by 14 hours. There is no threshold and no switch. There is a continuous shift in proportion that begins immediately.

That matters beyond pedantry, because the switch premise is what the various clock-hour claims elsewhere on this site are built on. If the underlying transition is gradual and already well advanced by hour 14, then a specific hour at which some downstream process begins has nothing to anchor to.

The second thing worth carrying is the protein cost. Protein loss during a fast is real, continuous, adaptively minimised and never abolished. Ketosis substantially reduces the rate compared with early fasting. It does not stop it. The 21 day study exists precisely to explain why death in starvation follows depletion of lean body mass while body fat still remains.