← Surfer · Sage hub

Sage — Obesity as Energy-Flux Management

A working hypothesis. Not medical advice, not a diagnosis, and not a reason to start, stop, or change a drug, a fast, or a diet. Prolonged fasting carries electrolyte imbalance, refeeding syndrome, and decompensation in frail or comorbid patients. It must be undertaken only under qualified medical supervision, with monitoring and a protocol that belongs to a clinician. This essay does not authorize a multi-day fast. GLP-1 receptor agonists have robust trial data for weight loss and for certain cardiovascular outcomes. They are not dismissed here. They are questioned for what they leave unaddressed.

The repair hypothesis: why medically supervised prolonged fasting with metabolic optimization offers a superior paradigm to GLP-1 agonists for obesity. The fat-loss hypothesis: insulin brake, spillover, depot order, oxidation.

Fat loss is release of stored triglyceride into mitochondrial oxidation. Insulin is the antilipolytic brake in every depot. Visceral and ectopic fat are spillover. A drug and a fast both work, when they work, by lowering insulin and imposing a deficit the mitochondria must meet. Neither is depot-selective.

Obesity, in the EnSurf frame, is an energy-flux problem of the whole system. The repair hypothesis already places the fast inside Fuel, Circadian, and Strength. Fuel is what arrives. Circadian timing is when it arrives and when light is allowed to set the repair window. Strength is capacity: mitochondrial volume, biogenesis, mitophagy, and whether the electron-transport chain can accept what beta-oxidation hands it. The Good Doctor’s three readings sit on that same system. Clean Line is ordinary operation — fuel, timing, and capacity coherent enough that fat is a buffer the engine can spend. Wipeout is the disease pattern — the buffer defended, then spilled, while the scale is mistaken for the engine. Kit is a modulation: a drug, a supervised fast, a named adjunct. Kit is not a fourth pillar of the ride. Foundation, if a reader needs the word, is the experiential field in which fuel, timing, and capacity are already happening. The three pillars are memory categories of those priors. They are not three systems, and Surfer, Coach, and Sage are not three medicines. Experience is the base. Nothing here requires a second substance called consciousness, and nothing here is an essay about ontology.

Dashboard override, and what it does not repair

The repair hypothesis opens in the incretin era and refuses the calorie-hack reading of a fast. The most powerful intervention, on its hypothesis, may be the one selection refined: not as a trick on intake, but as an engine recalibration. The dominant habit in chronic disease is to treat the dashboard — observable readouts and signaling overrides — and to neglect the engine: mitochondrial networks, autophagic machinery, insulin-signaling architecture, and the proteomic and epigenomic state that actually generates metabolic health or dysfunction. Obesity is not an exception.

GLP-1 receptor agonists — semaglutide, tirzepatide, and what has followed — are a sophisticated dashboard intervention. They suppress appetite through hypothalamic and brainstem signaling, slow gastric emptying, enhance glucose-dependent insulin secretion, and reduce glucagon. Large trials show clinically meaningful weight loss, often 15 percent or more of body weight, and, in selected populations, cardiovascular risk reduction. Those trial benefits are granted. They are not the whole claim.

The engine frequently stays compromised. Meta-analyses indicate roughly 25 to 39 percent of total weight lost on incretin therapy is lean mass, with semaglutide around 35 percent in some pooled data. That loss is described as comparable to diet-induced loss without resistance training. Adding a myostatin-pathway drug preserved substantially more lean mass in an early combination trial. On this hypothesis, though, that may be a forced phenotype: muscle kept by blocking a brake, without repairing the mitochondrial engine that powers it.

Maintenance typically requires ongoing pharmacotherapy. Discontinuation frequently leads to substantial regain. Gastrointestinal side effects drive discontinuation in a meaningful fraction of users. Long-term effects on muscle quality, bone, and other systems remain under investigation. On the repair hypothesis, the approach is, at root, chronic override of evolved control surfaces rather than repair of the bioenergetic machinery.

The different paradigm the repair hypothesis proposes is older than any drug: medically supervised prolonged fasting, typically five to seven or more days, water-only or with rigorous electrolyte support, integrated with targeted metabolic optimization before, during, and after. The claim is not that this merely unloads a broken engine. The claim is that it activates intrinsic repair programs that reconfigure the engine. That is the superiority hypothesis. It is not a home instruction, and it is not yet a verdict.

The day-3 threshold

The molecular spine of the repair hypothesis is a 2024 Nature Metabolism study, Pietzner, Uluvar, Kolnes, and colleagues, doi 10.1038/s42255-024-01008-9, “Systemic proteome adaptions to 7-day complete caloric restriction in humans,” Nature Metabolism 6: 764–777. Twelve healthy volunteers took a seven-day water-only fast. About 2,923 plasma proteins were tracked daily. Average weight loss was 5.7 kilograms. These were healthy volunteers, not a cohort selected for obesity. This essay does not promote the finding into a result already shown in patients with obesity.

Early changes, days 1 and 2, looked largely like the expected fuel switch as glycogen depleted. PCSK9 declined, which the repair hypothesis reads as consistent with shifting lipid flux.

The striking finding is the day-3 threshold. Of 1,034 significantly altered proteins, the majority did not move substantially until after day 3. From days 3 through 7 a second, larger wave of remodeling occurred — immune and inflammatory proteins, cellular stress responses, extracellular-matrix components from multiple tissues, including the brain-specific matrix protein tenascin-R, and pathways the repair hypothesis links to cardiovascular risk, autoimmunity, and longevity markers. By the end of the week the proteome had shifted widely from baseline, which the repair hypothesis reads as a new equilibrium. Many late changes mapped onto proteins and pathways that, in other literature the repair hypothesis invokes, may be associated with cardiovascular and autoimmune risk and with neuroprotection and longevity. That mapping is a hypothesis about meaning. It is not a claim that a seven-day fast has been shown to prevent those diseases.

Refeeding for three days largely reversed an apparent lean-mass loss of 3.6 kilograms during the fast, while fat-mass loss of about 1.85 kilograms was preserved. The repair hypothesis’s reading, not the paper’s, is that much of the “lean” signal in this healthy cohort reflected glycogen, water, and possibly liver glycogen or liver mass, rather than irreversible functional muscle catabolism.

Nick Norwitz’s commentary on the same work draws the practical line: many of the deepest, coordinated adaptations may require crossing the day-3 threshold. Shorter fasts may capture mainly the fuel switch. Prolonged fasts may unlock a broader systems recalibration.

The interpretation is engine work, not dashboard work. The body is not being forced into a lower-energy state by receptor agonism. It is running an evolved, multi-organ program of cleanup, fuel switching, stress-response recalibration, and structural remodeling that settles toward a new steady state. Those shifts are not side effects. On this hypothesis, they are the intervention.

Timing: a supervised fast used periodically, every several months or as needed, not a monthly rite.

Obesity as engine pathology

Obesity is rarely a primary disorder of willpower or of simple caloric arithmetic. On the repair hypothesis, it is often downstream of mitochondrial dysfunction and metabolic inflexibility: impaired oxidative capacity, fragmented mitochondrial networks, elevated reactive oxygen species with inadequate mitophagy, insulin resistance as both cause and consequence, and compensatory hyperphagia or fat storage driven by cellular energy-deficit signals. Earlier essays on atherosclerosis and on mitochondrial energy crisis are invoked as the same root. When the engine cannot convert fuel to ATP and cannot manage oxidative stress, the organism may compensate with behaviors and storage that look dysfunctional only on a dashboard.

GLP-1 agonism reduces the behavioral and glycemic symptoms of engine failure by overriding appetite and incretin signaling. It can lower the immediate load and buy time. It has not been shown, by itself, to restore mitochondrial biogenesis, enhance mitophagy of damaged organelles, or induce the proteomic and autophagic reprogramming seen once a fast passes day 3. In some analyses, the suppression of intake may itself limit the protein and micronutrients muscle needs during rapid loss — hence the interest in anti-myostatin adjuncts. Interest is not a recommendation.

The repair list:

AMPK activation and mTORC1 inhibition can trigger autophagy and mitophagy, clearing damaged proteins and organelles. This use of mTORC1 is the fasted, low-insulin state. It is not the same sentence as the claim that insulin, through PI3K, Akt, and mTORC1, inhibits adipose triglyceride lipase and thereby brakes lipolysis. Both can be true in different tissues at different moments. This essay does not fuse them into one switch.

Ketone bodies, beta-hydroxybutyrate in particular, are fuel and signal. The repair hypothesis names NLRP3 inhibition as the anti-inflammatory arm and histone-deacetylase inhibition as the epigenetic arm.

Hormonal shift: insulin declines, growth hormone rises, and late in a long fast FGF21 is induced. The repair hypothesis reads that pattern as promoting lipolysis while supporting signals that may preserve lean tissue and metabolic flexibility.

The post-day-3 proteome includes immune recalibration and stress-response tuning that may reduce the chronic low-grade inflammation the repair hypothesis treats as further engine damage.

On refeeding, a cleaned and partly rebuilt mitochondrial pool, plus anabolic signals from resistance training and protein coming back in, is hypothesized to support better functional recovery — and to partition incoming nutrients toward lean tissue rather than defaulting to fat storage. That partition is a hypothesis about suitable patients when resistance and protein are actually there. It is not a promise that any refeed does this.

Animal data further suggest that fasting periods are necessary for the full benefits of calorie restriction, and that continuous restriction without fasting windows can fail to turn on the same repair programs. No schedule is extracted from that sentence.

Diabetes remains its own page. In brief, as a proposed stack: lower PGC-1α and NRF-1; weaker electron-transport activity; reverse electron transport at complex I tied to liver coenzyme Q dysregulation; incomplete beta-oxidation with diacylglycerol and ceramides acting through PKCθ and JNK on IRS-1; seed-oil linoleic acid in cardiolipin and supercomplex assembly. Those lines explain a bottleneck the repair hypothesis names more generally as impaired oxidative capacity and insulin resistance. They do not replace the day-3 proteome, and they do not pull type 2 diabetes into this essay as a second disease to treat.

The hypothesis of fat loss

The fat-loss hypothesis starts one step closer to the lipid itself. Stored triglyceride is not a passive remainder and not a moral category. It is a buffer in a mitochondrial energy economy. In the fat-loss model, the cell oxidizes fatty acids when insulin is low enough to permit lipolysis and when the electron-transport chain can accept the acetyl-CoA that beta-oxidation makes. If either condition fails, the fatty acids do not vanish. They are re-esterified, redirected, or deposited where they do not belong. The public argument about fat loss has been held at the dashboard: scale, pinch, tape, weekly injection. The engine is upstream of all four.

The hypothesis: fat loss is release of stored triglyceride into mitochondrial oxidation, gated in both visceral and subcutaneous depots by insulin’s antilipolytic brake, and finished only if the receiving mitochondria can oxidize what was released. Visceral and ectopic fat are not a separate species. They are spillover. They appear when subcutaneous expandability fails and when cells cannot import or burn what leaks. Visceral adipocytes obey the same insulin brake as subcutaneous adipocytes, and they obey it less completely, which may be why they mobilize earlier as a percentage and why they may injure the liver on the way out. Subcutaneous fat, especially gluteofemoral fat, is the lower-risk warehouse. It is not a free good and it is not a toxin. GLP-1 receptor agonists and medically supervised prolonged fasting both work, when they work, by lowering insulin and forcing a deficit mitochondria must meet. Neither is a depot-selective instrument. The metabolic win is clearance of ectopic and visceral lipid with lean mass preserved and oxidative capacity restored. The scale is a late and unreliable witness.

Fat distribution is the spatial version of the failure the atherosclerosis, cancer, and aging essays described in time. A cell that cannot oxidize lipid stores it beside the work or ships it to a neighbor that also cannot oxidize it. The warehouse expands. The dashboard says obesity. The engine is still the mitochondrion.

The brake is not depot-specific

Lipolysis is three hydrolyses. Adipose triglyceride lipase cleaves triacylglycerol to diacylglycerol. Hormone-sensitive lipase cleaves diacylglycerol. Monoacylglycerol lipase releases glycerol and non-esterified fatty acids. Catecholamines raise cyclic AMP, activate protein kinase A, phosphorylate perilipin, free the ATGL coactivator CGI-58, and move hormone-sensitive lipase onto the droplet. That is the accelerator.

Insulin is the brake. At low physiological concentrations it activates phosphodiesterase 3B, lowers cyclic AMP, and withdraws the adrenergic signal. Through PI3K and Akt it also engages mTORC1, which inhibits ATGL. The same pathway is thought to run in visceral and subcutaneous adipocytes. There is no known private visceral lipase and no subcutaneous exemption. Fatty-acid release is exquisitely sensitive to insulin. That is why near-absence of insulin can run into ketoacidosis; the fat-loss hypothesis proposes that chronic hyperinsulinemia does the opposite: fatty acids locked in droplets while mitochondria elsewhere may be short of substrate.

Depots differ in gain, not in kind of machine. Visceral adipocytes are less sensitive to the brake and more sensitive to the accelerator: higher basal lipolysis, richer beta-adrenergic response, weaker insulin suppression. Abdominal subcutaneous adipocytes may sit between. Gluteofemoral adipocytes are the most insulin-sensitive and the slowest to release. Two people with the same total fat mass and the same fasting insulin may deliver very different fatty-acid loads to the liver, because the fraction of that mass behind a leaky brake is not the same.

Portal anatomy sharpens this without inventing a new biochemistry. Mesenteric and omental fat drain, in substantial part, into the portal vein, so fatty acids released there meet the hepatocyte before they meet skeletal muscle. Retroperitoneal fat is inside the abdomen and is often called visceral, but it is thought not to share that drainage. “Visceral” as one tissue is an imaging convenience. It is not one metabolic organ.

Spillover, not a second fat

The protective finding is not that subcutaneous fat is inert. It is that subcutaneous fat can expand. When preadipocytes recruit and new, smaller adipocytes hold surplus lipid, fatty acids are kept off liver, pancreas, myocardium, and skeletal muscle. Gluteofemoral subcutaneous tissue is the clearest buffer. Its expansion tracks a kinder insulin-sensitivity profile, and animal transplantation work has found that subcutaneous adipose tissue improved metabolic tone in mice even when moved into a visceral site. Visceral adipose tissue does not return the favor.

The pathogenic finding is failed expandability. Hypertrophic subcutaneous adipocytes that insulin can no longer restrain, or a failure to recruit new adipocytes, are proposed to spill lipid into visceral depots and into ectopic sites: hepatocyte droplets, intramyocellular lipid, pancreatic interstitial fat, epicardial fat. Visceral adipose tissue is itself a kind of ectopic storage. It is a minor depot, and it carries a disproportionate share of cardiometabolic risk, perhaps because it is leaky and portally drained. CT and MRI epidemiology finds it an independent marker of diabetes, atheroma, and cardiovascular events. Independence is real. It does not mean the subcutaneous warehouse was irrelevant. It means the warehouse failed.

Lipodystrophy is the counterexample to “adipose fat is harmless and only visceral fat is harmful.” Near-absence of subcutaneous storage produces severe insulin resistance, hepatic steatosis, and high triglycerides in people who are not obese on any scale. The missing buffer is the disease. Abdominal subcutaneous fat is not uniformly protective either. Large hypertrophic abdominal subcutaneous adipocytes associate with insulin resistance; hyperplastic expansion of the same depot may be less harmful. The variable is how the adipocyte behaves and where it sits, not the word adipose.

Absolute mass and percentage answer different questions, and the visceral-versus-subcutaneous argument usually drops one of them. Visceral fat is the smaller depot. Even when its percentage falls faster, most of the kilograms in any serious deficit still come from subcutaneous adipose tissue, because that is where most of the kilograms were. A tape that improves before the scale has moved much may often be percentage visceral loss plus the water that left with glycogen. It is not evidence that subcutaneous fat was spared.

Order of mobilization

In a sustained deficit or a fast, order follows lipolytic gain, blood supply, and insulin sensitivity. It is not a program that picks organs for moral reasons.

Liver glycogen goes first, over hours. Hepatic triglyceride is next among lipid stores and can fall substantially within a short period of carbohydrate restriction, because the hepatocyte is both a store and the organ that must meet gluconeogenic and ketogenic demand at once. Visceral adipose tissue follows on a percentage basis: higher basal lipolysis, denser catecholamine sensitivity, weaker insulin suppression, so it is the first large adipose depot to shrink as a fraction of itself. Subcutaneous adipose tissue is mobilized at the same time, more slowly as a percentage, and in greater absolute volume. Intramuscular and other small ectopic pools may fall with the deficit and add little to the mass change.

Animal fasting studies show the time dependence. Mesenteric fat is preferentially consumed relative to inguinal subcutaneous fat in the first day. Human imaging reviews find the same shape across interventions. At modest loss, the percentage drop in visceral fat often exceeds the percentage drop in subcutaneous fat. As loss deepens, the ratio approaches one. Preferential visceral loss is an early-phase description. It is not a promise that a long fast or a long drug course will leave subcutaneous fat in place.

The practical reading: waist and liver fat are leading indicators that the brake has lifted. The misleading reading is spot reduction. No diet, drug, or fast addresses a thigh while sparing a mesentery, or the reverse. Depots answer their receptors and their size. They do not take orders about a mirror.

Oxidation is what makes release into loss

A fatty acid freed from the droplet can be put back in the same adipocyte, taken up by the liver and rebuilt into VLDL, or oxidized. Insulin stimulates re-esterification even while it suppresses hydrolysis. Net export of non-esterified fatty acids therefore may fall further than glycerol release, because adipocytes are thought to have little glycerokinase, whereas fatty acids can be returned to the droplet. The brake has two hands.

The completing step is mitochondrial. Long-chain fatty acids enter the matrix on the carnitine shuttle, undergo beta-oxidation to acetyl-CoA, and feed the tricarboxylic acid cycle. Under low carbohydrate availability the liver converts surplus acetyl-CoA to acetoacetate and beta-hydroxybutyrate. Ketones are the transport form that lets adipose carbon reach brain, heart, and skeletal muscle without requiring a fatty-acid concentration in blood high enough to be toxic. Beta-hydroxybutyrate also signals, in animal and cell work: histone deacetylase inhibition and a lower inflammatory set-point; a PGC-1α-linked push toward biogenesis is proposed. Ketosis means lipolysis has exceeded the liver’s immediate oxidative demand and that insulin is low enough to permit the pathway. It does not, by itself, mean muscle mitochondria are competent.

Here the fat-loss hypothesis joins the fuel essay and the atherosclerosis essay on composition. Seed-oil polyunsaturated fatty acids in membranes and in the droplet become peroxidation substrates the moment lipolysis and rising reactive oxygen species coincide. Mobilizing a depot built from linoleic-acid-rich triglyceride is not the same act as mobilizing a depot built from saturated and monounsaturated fat. The aldehydes that follow — 4-hydroxynonenal, malondialdehyde — can adduct electron-transport subunits and, on this hypothesis, amplify the oxidative failure that forced storage in the first place. A fat-loss effort that ignores the composition of the fat being lost can improve the tape and worsen the redox account. That claim is the same seed-oil membrane problem the diabesity hypothesis placed inside cardiolipin.

Muscle is the sink that makes released fat safe. A myocyte with intact mitochondria, glycogen used and replaced by demand, and a creatine buffer recharged by oxidative phosphorylation may clear fatty acids that a sedentary, insulin-resistant myocyte will refuse. Refused fatty acids may return to the liver. A liver that cannot export or oxidize them becomes the next dashboard: steatosis, high triglyceride, hepatic insulin resistance. Fat loss that sacrifices the sink to empty the warehouse is, on this account, a bad trade. It may also be the common trade. The lean-mass cost named in the repair hypothesis is not a side-effect footnote. On the fat-loss account it is loss of the only large place the released fat can safely go.

GLP-1 agonism, three readings, bands not averaged

The diabesity hypothesis reports DXA analyses from pivotal trials in which lean mass makes up a widely varying share of the weight lost, and notes that a 15 kilogram loss can mean several kilograms of fat-free tissue, not all of it muscle. It names semaglutide 2.4 mg in the STEP program at about 15 percent mean loss at 68 weeks, tirzepatide higher, SELECT as an event-reduction trial against placebo, and SURPASS-CVOT as showing noninferiority to an active GLP-1 comparator. It notes that newer dual agonists may improve the ratio and that bimagrumab, given alone, cut fat and added lean mass in a trial. It offers a trajectory sketch, not a case: possible lean-mass loss in an older person with little reserve, with grip and aerobic capacity that may fall, and regain that may be disproportionately fat if the drug stops. None of those figures is averaged with the repair hypothesis or with the fat-loss hypothesis.

The fat-loss hypothesis cites a third band: across semaglutide and tirzepatide trials, something on the order of a fifth to two-fifths of weight lost is lean tissue unless resistance work and protein intake are deliberate. In the SURPASS-3 MRI substudy, tirzepatide reduced liver fat and both visceral and abdominal subcutaneous volumes beyond insulin degludec. Semaglutide series read on CT show paired decline in visceral and subcutaneous area among people who lose weight, liver volume down, liver attenuation up. Higher-dose semaglutide and dual and triple agonists produce larger absolute visceral losses. Relative visceral reductions appear in the imaging sets, and they track total weight loss. A 2025 meta-analysis of thirty-six randomized trials placed pooled visceral reduction with GLP-1-based drugs near 14 percent, behind diet plus exercise and behind very-low-energy diets. Fourteen percent visceral volume is not 15 percent scale weight and is not 25 to 39 percent lean mass.

Two inferences stay narrower than the advertising. The drugs are not visceral-selective. Visceral fat may shrink faster as a percentage because it is more lipolytically labile, which may be the same reason it shrinks faster in a fast. Subcutaneous fat still supplies most of the liters. The mechanism is thought to be mostly the deficit: less appetite, slower emptying, lower postprandial insulin demand, with effects on hepatic glucose production and, for dual agonists, on adipose insulin signaling. People who do not lose weight on the drug may not lose visceral fat. In at least one CT series, semaglutide users who gained weight gained visceral area, subcutaneous area, and intermuscular fat. A myostatin-pathway blocker added to a GLP-1 agonist can shift loss toward fat and has produced visceral reductions exceeding semaglutide alone with less lean-mass cost. That supports “preserve the sink,” not a hidden depot-targeting property, and it is not a protocol.

The repair hypothesis and the diabesity hypothesis agree, in different words, that the drug lowers load and does not rebuild beta-oxidation, biogenesis, or mitophagy. The fat-loss hypothesis agrees that the drug is a lever on insulin and intake. Discontinuation without a maintained deficit may return visceral and hepatic fat as the brake goes back on. The diabesity hypothesis’s additional cautions are possible downregulation of metabolic rate, gallbladder stasis, microbiome crosstalk, and cardiovascular benefit that may be the weight loss rather than mitochondrial repair of endothelium.

Good Doctor, only as far as it fits: a GLP-1 agonist can empty the plate and shrink the waist; lean mass often shrinks with the waist; the appetite signal is overridden rather than the engine repaired. That is Kit used as if it were the ride. Clean Line would be fuel, timing, and strength such that the brake lifts without a chronic clamp. Wipeout is celebrating the clamp because the dashboard moved.

The fast, read from the brake, not upgraded into superiority

The fat-loss hypothesis describes the same supervised fast from the adipocyte. A medically supervised fast drives insulin toward the floor, raises glucagon, growth hormone, and catecholamines, and shifts hepatic output from glycogen to gluconeogenesis and then to ketones. The antilipolytic brake is not modulated. It is released. Early mobilization may favor liver and visceral fat, though one 7-day human fast found the loss was mostly subcutaneous. Subcutaneous fat follows, and a long enough fast does not spare it. Ketones may spare some muscle relative to an unadapted low-calorie diet. They do not spare all of it. Growth hormone rises, perhaps in part to defend lean tissue and to permit lipolysis. It is a compensating signal, not a guarantee. The repair hypothesis also has insulin down and growth hormone up, and adds FGF21. The fat-loss hypothesis does not name FGF21.

On the fat-loss hypothesis’s comparison, the advantage of the fast over the drug is depth, and the disadvantage is the same depth. Insulin is lower. Ketogenesis is fuller. Autophagy and mitophagy may have a clearer window. Sodium, potassium, magnesium, and phosphate can leave with glycogen water and with the natriuresis of low insulin. Refeeding after depletion can precipitate the syndrome named for it. A long fast in a person whose muscle is already poor, whose depot is rich in peroxidizable polyunsaturated fat, or whose medications assume a fed insulin level is not a cleaner GLP-1. It is a larger perturbation of the same axis, with a narrower safety margin. Neither tool is a chronic strategy by itself. The fast is a depot-clearing interval. The drug is a pharmacological interval of reduced intake. Both fail if the weeks after them restore hyperinsulinemia, seed-oil membranes, and disuse. Both succeed if ectopic fat stays down, visceral percentage stays down, and the capacity to oxidize fatty acids is higher at the end than at the start.

Optimization, without amounts

Bare prolonged fasting is already powerful, and optimization consistent with Fuel, Circadian, and Strength amplifies engine repair. No milligram, milliliter, or packet size appears in that sketch, and none is added here. Jason Fung and Jimmy Moore, The Complete Guide to Fasting, are cited as a practical starting point for extended-fasting protocols, safety considerations, and clinical rationale, with the proteomic data as the reason duration matters. This essay does not turn it into a schedule.

Before the fast, the repair hypothesis names mitochondrial support as classes it proposes: PQQ (proposed for biogenesis), coenzyme Q10, and perhaps low-dose methylene blue or NAD precursors depending on individual status. “Low-dose” and “depending on status” are qualifiers. They are not a dose. Mineral repletion and short conditioning fasts are named to improve tolerance. Amounts are not stated.

During the fast: electrolyte management named as classes — sodium, potassium, magnesium, phosphate — with monitoring, not as a chart; adequate hydration; circadian light exposure; rest with light movement. Heroic exercise that could accelerate excessive catabolism is to be avoided. It is not a training plan.

After the fast: careful, phased reintroduction; nutrient density from animal proteins and fats as bioavailable building blocks; fermented microbiota-accessible carbohydrates for microbiome recolonization; creatine to support ATP buffering and muscle; resistance training, dynamic or ARX-style, as the mechanical and metabolic signal for anabolism on a cleaned mitochondrial platform; continued mitochondrial support and circadian discipline. The sequence is hypothesized to turn the fast from a stress into a hormetic cycle of clearance and then reconstruction — the difference between simply losing weight and emerging with upgraded metabolic hardware.

The fat-loss hypothesis adds fuel and timing lines. Industrial seed oils are treated there as the composition of the droplet being mobilized, not only as a vague processed-food category. Animal fat and protein as dominant substrates are proposed there to keep beta-oxidation in its preferred mode. An overnight interval that ends eating several hours before sleep is an ordinary way to lower insulin before a drug or a fast. “Several hours” is not converted here into a clock time or a 16:8 rule. Resistance work, heavy and infrequent enough to recover from, is proposed to defend the sink. Zone-2 movement spends fatty acids without the cortisol cost of chronic high intensity. It is not given as a heart-rate formula. Fermented foods and resistant starches match fermented microbiota-accessible carbohydrates; they are not a second diet.

The diabesity hypothesis’s adjunct list — creatine, PQQ, coenzyme Q10, timed NAD precursors — overlaps the class names and adds nothing that should be dosed. Its claim that morning light, meal timing, and darkness are among the highest-leverage interventions is the circadian pillar said again, not a new protocol.

Durability, and the frugal contrast

Dashboard tools excel at rapid loss on the scale. Engine tools prioritize what is lost and what remains after the intervention ends. GLP-1 lean-mass loss is called clinically relevant especially in older adults or people with low baseline muscle. Resistance training improves the ratio. The catabolic pressure of a sustained deficit, plus possible interference with protein intake, remains.

In the proteomics study, much of the lean signal reversed quickly with refeeding, and fat loss persisted. The repair hypothesis then points at clinical therapeutic-fasting settings, including protocols informed by Jason Fung’s work and similar supervised programs, where patients often report preserved or improved strength after refeed when resistance training and protein reintroduction are emphasized. “Often report” is clinical impression, not the Pietzner endpoint. The engine, once cleared of damaged components and shifted toward fat oxidation, is hypothesized to partition incoming nutrients toward lean tissue rather than defaulting to fat storage.

Durability follows, still as hypothesis. GLP-1 weight loss is largely maintained only while the drug is continued, for most patients. Periodic or strategic prolonged fasting is proposed as a reset under supervision, every several months or as needed, while daily metabolic optimization — circadian alignment, resistance training, mitochondrial support, nutrient-dense refeeding — maintains the gains. The cost contrast is a claim and not a price list: a supervised fast is argued to be far less expensive than years of GLP-1 therapy at current pricing. Frugality, in that paragraph, is not cheapness. It is alignment with high-leverage evolved mechanisms that are argued to compound as resilience, rather than a perpetual subscription to pharmacological signaling. This essay invents no dollar figure on either side.

What would count, and what reaches past the scale

The tests, falsifiable and still undone: body composition by DEXA or an advanced method, before and after; fasting insulin and HOMA-IR; hs-CRP; an oxidative-stress marker such as 8-OHdG; functional metrics, grip strength; targeted proteomics or metabolomics where accessible. Durability at 6 and 12 months after the intervention, set against matched GLP-1 or lifestyle cohorts. Head-to-head or pragmatic trials in appropriate populations, with supervision standards and refeeding protocols attended to, not assumed.

The hypothesis: for suitable patients, under rigorous medical supervision and integrated with targeted optimization, medically supervised prolonged fasting more effectively restores engine function — and therefore sustainable metabolic health — than chronic GLP-1 agonism alone. The pointers it stacks are the duration-dependent proteomic remodeling, alignment with autophagy, mitophagy, and metabolic flexibility, the body-composition and durability considerations, and a preference for root causes over surface controls.

The fat-loss hypothesis’s success readings stay beside those tests and are not swapped in for them: hepatic fat; visceral volume as a percentage read next to absolute subcutaneous volume, so the buffer is not mistaken for the disease; the insulin concentration required to suppress non-esterified fatty-acid flux; respiratory quotient under a standard challenge; lean mass, grip, and clearance of a fatty-acid load into working muscle; ketone appearance on an ordinary overnight fast. A loss of gluteofemoral fat with a gain in liver fat can look like progress and be a deterioration. A stable scale with falling liver fat can look like failure and be a repair.

Beyond weight, the repair hypothesis treats obesity as a gateway and amplifier for other engine-level pathologies — atherosclerosis, as hypothesized elsewhere in metabolic terms; insulin resistance driving further mitochondrial stress; chronic inflammation; vulnerability that may extend further. An intervention that reduces adiposity and also recalibrates immune tone, stress responses, and bioenergetic efficiency is argued to matter past the scale. Coordinated multi-organ shifts after day 3, including brain extracellular matrix and tenascin-R, are offered as effects dashboard weight-loss drugs were not designed to produce. That is a hypothesis about reach. It is not a claim of neuroprotection proven in this fast.

Where the repair hypothesis and the fat-loss hypothesis do not agree

They agree on the engine. Obesity is not a primary failure of character, and it is not solved by a prettier scale. The mitochondrion is upstream. GLP-1 agonism is a real lever on appetite and on insulin, with real trial benefits, and it is not, by itself, a repair of biogenesis and mitophagy. Regain after the lever is removed is expected if nothing in the machinery was rewritten. A prolonged fast, if it is used at all, is supervised and time-limited, and it carries electrolyte and refeeding risk that a webpage cannot manage. Muscle matters. A loss that spends the sink to empty the warehouse fails the test both care about, even though only the fat-loss hypothesis says “sink” in those words.

They agree on depot order, because that order is the fat-loss hypothesis’s contribution. Visceral and liver fat move early as percentages. Subcutaneous fat supplies most of the kilograms. Spot reduction is a myth. Lipodystrophy shows that a missing subcutaneous buffer is itself a disease. The repair hypothesis does not make those depot distinctions. Importing them does not make the repair hypothesis depot-selective, and the fat-loss hypothesis denies that either tool is.

They do not agree on what “superior” means, and this essay does not pick a winner.

The repair hypothesis is that a supervised prolonged fast plus metabolic optimization restores engine function better than chronic GLP-1 agonism alone, because it crosses a day-3 proteomic threshold, turns on AMPK-linked autophagy and mitophagy, uses beta-hydroxybutyrate as an NLRP3 and histone signal, shifts insulin, growth hormone, and FGF21, and — when refeeding includes resistance and protein — partitions toward lean tissue. Superiority there is about repair and durability, not about a larger week-one kilogram loss. The proteomics study it relies on was in twelve healthy volunteers. The clinical step to people with obesity is the hypothesis, not a result already in that paper.

The fat-loss hypothesis is stingier. It grants the fast a deeper signal — lower insulin, fuller ketogenesis, a clearer autophagy and mitophagy window — and prices that depth as danger. It predicts that GLP-1 agonists and prolonged fasts will not differ in depot order once insulin and energy deficit are matched. They will differ in ketosis, growth-hormone tone, lean-mass cost, and how fast fat returns after the lever is removed. It also says outright that neither tool is a chronic strategy. The repair hypothesis’s “every several months or as needed,” maintained by daily optimization, is a repeated reset, not a chronic fast, but it is more willing than the fat-loss hypothesis to treat the fast as the superior paradigm rather than as one interval among others.

A tension: repair-superiority versus same-axis, greater-amplitude, narrower-margin. Both can be true if “repair” is real and if depot order is still just lipolytic gain. Both cannot be true if “superior” is rewritten to mean “more kilograms from the waist, safely, at home.” This essay does not say that.

If a future study matches insulin and deficit and still finds a proteomic or mitochondrial difference that favors the fast, the repair hypothesis’s claim gains a point the fat-loss hypothesis did not spend. If matched deficits produce the same depots, the same lean-mass fraction, and the same mitochondrial markers, the superiority claim fails and the fat-loss hypothesis’s reading is what remains.

Predictions

From the repair hypothesis, not yet run as the trials it asks for:

• In suitable patients under real supervision, prolonged fasting plus the optimization sequence above restores engine readouts — composition, fasting insulin and HOMA-IR, hs-CRP, an oxidative marker such as 8-OHdG, grip — more effectively than chronic GLP-1 agonism alone, and the difference is still visible at 6 and 12 months.
• Much of the early lean-mass signal of a supervised water fast reverses with refeeding, as in the healthy-volunteer proteomics study; the prediction is that this holds when the refeed is phased and includes protein and resistance. That study’s 3.6 kilogram lean loss, largely reversed, and 1.85 kilogram persistent fat loss are that cohort’s numbers, not a quota.

From the fat-loss hypothesis:

• People who lower fasting insulin will show an early percentage decline in visceral and hepatic fat that exceeds the percentage decline in gluteofemoral fat, while absolute mass loss is still dominated by subcutaneous adipose tissue.
• Interventions that preserve or increase lean mass will produce more durable reductions in hepatic fat than matched weight loss that sacrifices muscle.
• Depots enriched in polyunsaturated seed-oil fat will show more lipid-peroxidation products during rapid mobilization than depots enriched in saturated and monounsaturated fat, at the same rate of loss.
• GLP-1 agonists and prolonged fasts will not differ in depot order once insulin and energy deficit are matched. They will differ in ketosis, growth-hormone tone, lean-mass cost, and speed of regain after the lever is removed.
• Lipodystrophy and failed subcutaneous expandability will keep producing the visceral and ectopic picture without obesity.

From the diabesity hypothesis:

• Long-term GLP-1 monotherapy without resistance training and a change in dietary substrate will show more sarcopenia, frailty, and rebound after cessation than protocols aimed at mitochondrial restoration.
• Mitochondrial markers will improve more with fasting, resistance loading, and polyunsaturated-fat reduction than with GLP-1 agonism alone, even when the drug produces greater short-term weight loss.
• High lifetime seed-oil exposure will worsen the lean-to-fat loss ratio on GLP-1 agents.
• Circadian and resistance work, with or without a GLP-1, will maintain weight, A1c, and function better than pharmacotherapy alone.

The second of those diabesity predictions rhymes with the repair hypothesis and is not identical to it. The repair hypothesis’s comparator is chronic GLP-1 alone, in suitable supervised patients, on engine function and durability. The diabesity line adds seed-oil reduction as a named co-intervention and allows the drug to win on short-term weight.

What this essay refuses

It does not deny trial benefits of GLP-1 receptor agonists for weight and for selected cardiovascular outcomes. It does not deny surgery or other indicated care. It does not claim that an unsupervised fast, an extreme diet, or a supplement reverses obesity or replaces obesity medicine. It does not tell anyone to start or stop a GLP-1 or any other drug. It does not invent a fasting calendar, a price, an electrolyte chart, or a refeeding menu. Fung and Moore are cited, not rewritten as instructions.

It does not claim anyone manufactures a product, and it does not sell one. It does not collapse the diabetes essay into this obesity essay. Coach and Surfer compress the same hypothesis. They are not a second medicine. Diagnosis, medication clocks, and any decision about a prolonged fast stay with the person and a clinician who knows the case.

Obesity is not fundamentally a shortage of GLP-1 signaling. It is often a failure of the engine to handle energy, clear damage, and stay flexible. GLP-1 receptor agonists have a place for people who need rapid risk reduction or who have not responded to other approaches. They manage a dashboard while, on this hypothesis, the machinery is not repaired. A supervised fast with metabolic optimization is argued to do something different: conserved repair programs at the proteomic, autophagic, mitochondrial, and hormonal levels, past a day-3 threshold where the remodeling changes kind, at a cost the repair hypothesis calls frugal because it borrows evolved biology instead of overriding it indefinitely.

Stored fat is the buffer. Visceral fat is overflow when the brake cannot hold. Subcutaneous fat is not the enemy and is not innocent. Fat loss is a permission — insulin down, lipase on, carnitine shuttle intact, muscle willing to burn what the depot releases — and a fast or a drug only borrows a signal the organism already uses when the night is actually fasted. The repair hypothesis has not yet granted the fat-loss hypothesis’s refusal of repair-superiority. This essay will not grant it either by silence.

GLP-1 therapy can serve as a bridge for urgent need. The destination, in this hypothesis, is restoration of autonomous metabolic integrity — not indefinite receptor agonism mistaken for healed terrain.

Limits of the claim

• Working hypothesis. Not a randomized-trial verdict that mitochondrial-first strategies outperform guideline obesity care as monotherapy.
• Does not deny trial benefits of GLP-1 RAs for weight and selected cardiovascular outcomes, nor the clinical utility of surgery or other indicated therapies.
• Does not claim unsupervised prolonged fasting, extreme diets, or supplements reverse obesity or replace obesity-medicine clinics.
• Diabetes has its own page (sequestration / “water, water everywhere”). Diabetes and obesity share one engine but have separate pages.
• No dualism.

Coach and Surfer give shorter versions. Obesity medicine and primary care still own diagnosis, medication clocks, and any decision about prolonged fasting.

A working hypothesis, open to correction and revised as new evidence arrives. Last revised October 31, 2026.

References

1. Wilding JPH, Batterham RL, Calanna S, et al., 2021. Once-Weekly Semaglutide in Adults with Overweight or Obesity. New England Journal of Medicine 384(11):989-1002. doi:10.1056/NEJMoa2032183
2. Lincoff AM, Brown-Frandsen K, Colhoun HM, et al., 2023. Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes. New England Journal of Medicine 389(24):2221-2232. doi:10.1056/NEJMoa2307563
3. Eisa N, Barood O, 2026. Lean Mass Changes With Incretin Therapy Versus Lifestyle Intervention: A Systematic Review and Meta-Analysis of Randomised Controlled Trials. Diabetes, Obesity and Metabolism 28(6):4818-4827. doi:10.1111/dom.70666
4. Wilding JPH, Batterham RL, Davies M, et al., 2022. Weight regain and cardiometabolic effects after withdrawal of semaglutide: The STEP 1 trial extension. Diabetes, Obesity and Metabolism 24(8):1553-1564. doi:10.1111/dom.14725
5. Pietzner M, Uluvar B, Kolnes KJ, et al., 2024. Systemic proteome adaptions to 7-day complete caloric restriction in humans. Nature Metabolism 6(4):764-777. doi:10.1038/s42255-024-01008-9
6. Xie Z, Zhang D, Chung D, et al., 2016. Metabolic Regulation of Gene Expression by Histone Lysine β-Hydroxybutyrylation. Molecular Cell 62(2):194-206. doi:10.1016/j.molcel.2016.03.036
7. Kim J, Kundu M, Viollet B, et al., 2011. AMPK and mTOR regulate autophagy through direct phosphorylation of Ulk1. Nature Cell Biology 13(2):132-141. doi:10.1038/ncb2152
8. Chakrabarti P, Kim JY, Singh M, et al., 2013. Insulin inhibits lipolysis in adipocytes via the evolutionarily conserved mTORC1-Egr1-ATGL-mediated pathway. Molecular and Cellular Biology 33(18):3659-3666. doi:10.1128/MCB.01584-12
9. Shimazu T, Hirschey MD, Newman J, et al., 2013. Suppression of oxidative stress by β-hydroxybutyrate, an endogenous histone deacetylase inhibitor. Science 339(6116):211-214. doi:10.1126/science.1227166
10. Youm YH, Nguyen KY, Grant RW, et al., 2015. The ketone metabolite β-hydroxybutyrate blocks NLRP3 inflammasome-mediated inflammatory disease. Nature Medicine 21(3):263-269. doi:10.1038/nm.3804
11. Ho KY, Veldhuis JD, Johnson ML, et al., 1988. Fasting enhances growth hormone secretion and amplifies the complex rhythms of growth hormone secretion in man. Journal of Clinical Investigation 81(4):968-975. doi:10.1172/JCI113450
12. Fazeli PK, Lun M, Kim SM, et al., 2015. FGF21 and the late adaptive response to starvation in humans. Journal of Clinical Investigation 125(12):4601-4611. doi:10.1172/JCI83349
13. Pak HH, Haws SA, Green CL, et al., 2021. Fasting drives the metabolic, molecular and geroprotective effects of a calorie-restricted diet in mice. Nature Metabolism 3(10):1327-1341. doi:10.1038/s42255-021-00466-9
14. Duncan RE, Ahmadian M, Jaworski K, et al., 2007. Regulation of lipolysis in adipocytes. Annual Review of Nutrition 27:79-101. doi:10.1146/annurev.nutr.27.061406.093734
15. Jensen MD, Nielsen S, 2007. Insulin dose response analysis of free fatty acid kinetics. Metabolism 56(1):68-76. doi:10.1016/j.metabol.2006.08.022
16. Ibrahim MM, 2010. Subcutaneous and visceral adipose tissue: structural and functional differences. Obesity Reviews 11(1):11-18. doi:10.1111/j.1467-789X.2009.00623.x
17. Meek SE, Nair KS, Jensen MD, 1999. Insulin regulation of regional free fatty acid metabolism. Diabetes 48(1):10-14. doi:10.2337/diabetes.48.1.10
18. Manolopoulos KN, Karpe F, Frayn KN, 2010. Gluteofemoral body fat as a determinant of metabolic health. International Journal of Obesity 34(6):949-959. doi:10.1038/ijo.2009.286
19. Tran TT, Yamamoto Y, Gesta S, et al., 2008. Beneficial effects of subcutaneous fat transplantation on metabolism. Cell Metabolism 7(5):410-420. doi:10.1016/j.cmet.2008.04.004
20. Fox CS, Massaro JM, Hoffmann U, et al., 2007. Abdominal visceral and subcutaneous adipose tissue compartments: association with metabolic risk factors in the Framingham Heart Study. Circulation 116(1):39-48. doi:10.1161/CIRCULATIONAHA.106.675355
21. Garg A, 2011. Clinical review#: Lipodystrophies: genetic and acquired body fat disorders. Journal of Clinical Endocrinology and Metabolism 96(11):3313-3325. doi:10.1210/jc.2011-1159
22. Weyer C, Foley JE, Bogardus C, et al., 2000. Enlarged subcutaneous abdominal adipocyte size, but not obesity itself, predicts type II diabetes independent of insulin resistance. Diabetologia 43(12):1498-1506. doi:10.1007/s001250051560
23. Browning JD, Baker JA, Rogers T, et al., 2011. Short-term weight loss and hepatic triglyceride reduction: evidence of a metabolic advantage with dietary carbohydrate restriction. American Journal of Clinical Nutrition 93(5):1048-1052. doi:10.3945/ajcn.110.007674
24. Tang HN, Tang CY, Man XF, et al., 2017. Plasticity of adipose tissue in response to fasting and refeeding in male mice. Nutrition & Metabolism 14:3. doi:10.1186/s12986-016-0159-x
25. Chaston TB, Dixon JB, 2008. Factors associated with percent change in visceral versus subcutaneous abdominal fat during weight loss: findings from a systematic review. International Journal of Obesity 32(4):619-628. doi:10.1038/sj.ijo.0803761
26. Nicholls SJ, Pavo I, Bhatt DL, et al., 2025. Cardiovascular Outcomes with Tirzepatide versus Dulaglutide in Type 2 Diabetes. New England Journal of Medicine 393(24):2409-2420. doi:10.1056/NEJMoa2505928
27. Heymsfield SB, Aronne LJ, Montgomery P, et al., 2026. Bimagrumab plus semaglutide alone or in combination for the treatment of obesity: a randomized phase 2 trial. Nature Medicine 32(3):869-882. doi:10.1038/s41591-026-04204-0
28. Karakasis P, Patoulias D, Fragakis N, et al., 2025. Effect of glucagon-like peptide-1 receptor agonists and co-agonists on body composition: Systematic review and network meta-analysis. Metabolism 164:156113. doi:10.1016/j.metabol.2024.156113
29. Gastaldelli A, Cusi K, Fernández Landó L, et al., 2022. Effect of tirzepatide versus insulin degludec on liver fat content and abdominal adipose tissue in people with type 2 diabetes (SURPASS-3 MRI): a substudy of the randomised, open-label, parallel-group, phase 3 SURPASS-3 trial. The Lancet Diabetes & Endocrinology 10(6):393-406. doi:10.1016/S2213-8587(22)00070-5
30. Nelson LW, Lee MH, Garrett JW, et al., 2024. Intrapatient Changes in CT-Based Body Composition After Initiation of Semaglutide (Glucagon-Like Peptide-1 Receptor Agonist) Therapy. AJR American Journal of Roentgenology 223(6):e2431805. doi:10.2214/AJR.24.31805
31. Ghosal S, Sinha B, Ghosal A, et al., 2025. Comparative efficacy of lifestyle and GLP-1RA-based interventions for visceral adipose tissue reduction: meta-analysis of 36 randomised controlled trials. Obesity Abstracts 5 (Obesity Update 2025), P1. doi:10.1530/obabs.05.p1
32. Mehanna HM, Moledina J, Travis J, 2008. Refeeding syndrome: what it is, and how to prevent and treat it. BMJ 336(7659):1495-1498. doi:10.1136/bmj.a301
33. Drucker DJ, 2018. Mechanisms of Action and Therapeutic Application of Glucagon-like Peptide-1. Cell Metabolism 27(4):740-756. doi:10.1016/j.cmet.2018.03.001
34. Virtue S, Vidal-Puig A, 2010. Adipose tissue expandability, lipotoxicity and the Metabolic Syndrome--an allostatic perspective. Biochimica et Biophysica Acta 1801(3):338-349. doi:10.1016/j.bbalip.2009.12.006
35. Britton KA, Massaro JM, Murabito JM, et al., 2013. Body fat distribution, incident cardiovascular disease, cancer, and all-cause mortality. Journal of the American College of Cardiology 62(10):921-925. doi:10.1016/j.jacc.2013.06.027
36. Hoffstedt J, Arner E, Wahrenberg H, et al., 2010. Regional impact of adipose tissue morphology on the metabolic profile in morbid obesity. Diabetologia 53(12):2496-2503. doi:10.1007/s00125-010-1889-3
37. Puchalska P, Crawford PA, 2017. Multi-dimensional Roles of Ketone Bodies in Fuel Metabolism, Signaling, and Therapeutics. Cell Metabolism 25(2):262-284. doi:10.1016/j.cmet.2016.12.022

See the notice at the top of this page. Working hypothesis, not medical advice. Essay draws from grounded mechanisms in mitochondrial biology, obesity metabolism, fasting proteomics, and related clinical observation as of 2026 — individual care requires personalized evaluation by a qualified physician.