Nutrition · Pro: every answer, checked

The indicator amino acid oxidation (IAAO) method produced requirement estimates 30-50% higher than the nitrogen balance studies that older guidelines were built on. IAAO puts recommended intakes near 1.6-1.8g/kg/day for endurance athletes and up to 2.2g/kg/day for strength athletes. Because nitrogen balance systematically underestimates requirements, the field shifted toward the stable-isotope estimates.

Nitrogen balance overestimates intake and underestimates losses through skin, sweat, and miscellaneous routes, which biases results toward apparent balance at low intakes. The urea pool also adapts slowly to a new protein intake, so short study periods capture incomplete adaptation. Together these flaws push requirement estimates downward.

IAAO testing in endurance-trained men, performed on the day of a 20km run within a controlled training block, found protein requirements exceeding both the RDA and the classic 1.2-1.4g/kg athlete guideline. Estimated average requirement landed near 1.65g/kg/day with a recommended intake around 1.83g/kg/day. The finding held even though endurance training was long assumed to spare protein through adaptation.

Sestrin2 is a direct leucine sensor. In leucine's absence, Sestrin2 binds and inhibits GATOR2, an upstream activator of mTORC1. Leucine binds Sestrin2 with a dissociation constant near 20 micromolar, triggering a conformational change that releases GATOR2, which then permits Rag GTPase-dependent mTORC1 activation at the lysosome.

No. Mitchell and colleagues (2014) found acute post-exercise myofibrillar FSR in novices was uncorrelated with quadriceps growth after chronic training, with correlation coefficients near zero across 1-6 hour measurement windows. Acute MPS responses are a measure of the momentary stimulus, not a proxy for the training outcome.

An infusion protocol samples a window of a few hours in a lab-controlled state, while the post-exercise synthetic response extends over 24-72 hours and interacts with sleep, meals, and subsequent activity. Extrapolating from the short window assumes it represents every later response, which the data do not support. Muscle protein breakdown is also rarely measured alongside it.

D2O is drunk orally and labels body water for days to weeks, so it measures integrated MPS under free-living conditions: meals, training, sleep, and daily activity all included. That captures the summed anabolic and catabolic reality that a 4-12 hour intravenous infusion in a lab cannot. The trade is less control over acute conditions.

Evidence points to a cluster: reduced muscle blood flow and microvascular perfusion limiting amino acid delivery, decreased amino acid uptake capacity, and blunted mTORC1 activation within the myocyte despite adequate circulating amino acids. The practical result is that older muscle needs a larger protein and leucine dose to mount the same synthetic response.

DIAAS uses true ileal digestibility of each individual amino acid, measured at the end of the small intestine. PDCAAS uses fecal digestibility of total nitrogen, which is confounded by hindgut microbial metabolism and blind to differences between amino acids. Ileal, amino-acid-specific values are considered the more accurate basis.

PDCAAS is truncated at 1.0, so every protein at or above the reference pattern gets the same capped score. Whey and soy isolates both sit near 1.0 under PDCAAS. DIAAS is uncapped, so whey scores above 100 while soy commonly lands around 75-90, exposing a quality gap the truncation hides.

Collagen is an incomplete protein: it lacks tryptophan and is low in leucine, the amino acid that gates mTORC1 activation. Trials adding collagen to resistance training show no additive effect on muscle strength or mass beyond training alone. Its proposed use is connective tissue support, not muscle anabolism.

The idea comes from Shaw and colleagues (2017), where 15g of vitamin C-enriched gelatin before intermittent activity roughly doubled a blood marker of collagen synthesis in eight men. Later work is mixed: some trials adding 5-15g collagen to loading programs report tendon structural improvements, but clinical outcome evidence remains sparse, heterogeneous, and low certainty. Promising mechanism, unproven clinical claim.

The 1960s model demanded 3 days of exhaustive exercise and near-zero carbohydrate before loading. Later work showed trained athletes have chronically higher glycogen synthase activity and greater storage capacity, so they reach supercompensated glycogen with high carbohydrate intake and taper alone. The depletion phase added fatigue and irritability without adding glycogen.

Roughly 10-12g/kg/day of carbohydrate for 24-48 hours before the event, combined with rest or minimal training. In trained athletes this achieves muscle glycogen levels comparable to the week-long classic protocol. Events under about 90 minutes generally do not need a formal load at all.

Muscle contraction triggers GLUT4 translocation to the cell membrane through an insulin-independent pathway involving AMPK and calcium signaling. This is additive to the insulin pathway, so during and immediately after exercise glucose enters muscle at high rates even with low insulin. It is one reason post-exercise carbohydrate is stored so efficiently.

Intestinal absorption, not muscle. The sodium-dependent glucose transporter SGLT1 in the gut brush border saturates around 60g/h of glucose intake. Beyond that, extra glucose sits in the gut, raising GI distress risk without raising oxidation, which plateaus near 1g/min.

Fructose crosses the gut via GLUT5, a separate transporter from the SGLT1 route glucose uses. Combining roughly 60g/h glucose with 30g/h fructose runs both transporters in parallel, lifting peak exogenous oxidation from about 1g/min with glucose alone to around 1.26g/min, and up to 1.75g/min at the highest mixture intakes studied. Total delivery rises without overloading either pathway.

A chronically high carbohydrate intake, including carbs during training, is thought to raise intestinal absorption capacity via SGLT1 upregulation, lifting exogenous oxidation and lowering GI symptom rates. Cox and colleagues (2010) showed higher exogenous carbohydrate oxidation after 28 days of high carbohydrate availability. Evidence is moderate: oxidation data are solid, direct transporter measurements come mostly from animal work, and performance trials are fewer.

After about 3 weeks on a ketogenic low-carb high-fat diet, fat oxidation rates rose dramatically, but the oxygen cost of walking at race speeds increased, meaning worse exercise economy. High-carb groups improved race times after the training block; the LCHF group did not, despite similar VO2peak gains. The result was later replicated by the same group.

Oxidizing fat yields less ATP per liter of oxygen consumed than oxidizing carbohydrate, a gap of roughly 5-8%. At low intensity with spare aerobic capacity this is irrelevant. Near competitive intensities, where oxygen delivery is the ceiling, the extra O2 demand per watt makes fat-adapted athletes less economical.

Training with low glycogen amplifies cell signaling: greater AMPK activity, higher PGC-1alpha expression, and larger increases in oxidative enzymes. The performance translation is weak, with a minority of studies showing improvement and meta-analysis of periodized carbohydrate restriction finding no overall benefit versus normal fueling. Signaling endpoints and race results are not the same outcome.

Only when carbohydrate intake is suboptimal, around 0.8g/kg/h or less. There, added protein raises insulin and improves glycogen storage rates. When carbohydrate is already at the maximizing dose of about 1.2g/kg/h, protein co-ingestion adds nothing further to glycogen. Protein still serves muscle repair either way.

Absorbed fructose is largely cleared on first pass by the liver, where fructokinase traps it as fructose-1-phosphate, bypassing the rate-limiting phosphofructokinase step. Skeletal muscle expresses little capacity to take up and use fructose directly. So glucose-fructose mixtures after exercise restore liver glycogen faster than glucose alone while muscle repletion runs on the glucose fraction.

Creatine kinase transfers a phosphate between phosphocreatine and ADP, regenerating ATP within milliseconds at the site of contraction, and shuttles high-energy phosphate between mitochondria and myofibrils. Supplementation raises muscle total creatine roughly 20% and phosphocreatine around 10%, extending the capacity of this buffer during repeated maximal efforts and speeding PCr resynthesis between sets.

Vegetarians, whose baseline muscle creatine runs 20-30% below meat eaters, show the largest gains in total creatine and phosphocreatine when supplementing. An estimated 20-30% of users are non-responders, typically because their stores are already near the saturation ceiling. Response size tracks the gap between baseline and maximal muscle creatine content.

Caffeine is an adenosine receptor antagonist. Adenosine accumulates with sustained effort and signals fatigue by binding A1 and A2A receptors in the central nervous system; caffeine occupies those receptors without activating them. The dominant result is reduced perception of effort and pain, with secondary effects on motor unit recruitment. The old fat-sparing glycogen theory is no longer considered the main driver.

The evidence is contested. Some studies find AA (fast metabolizer) genotypes benefit most, with smaller or null effects in AC and CC carriers, but other trials find similar ergogenic responses across genotypes. Systematic reviews conclude there is no consensus and the effects are highly variable. Genotype-based caffeine prescriptions are ahead of the data.

No. Goncalves and colleagues (2017) found the performance response to acute caffeine in a cycling time trial did not differ across low, moderate, and high habitual consumers. Later reviews broadly support that habitual intake does not eliminate the acute benefit. Withdrawal protocols before competition are therefore not required, though some athletes still prefer them.

Exercise raises IL-6, which stimulates hepatic hepcidin release peaking around 3-6 hours post-session. Hepcidin degrades ferroportin, the transporter that exports iron from gut cells and macrophages into blood, so iron taken in that window is poorly absorbed. This inflammatory loop is one reason endurance athletes run higher iron deficiency rates despite adequate intakes.

In the morning, and at least several hours away from training. Hepcidin shows diurnal variation, lowest in the early morning and rising through the day, and it also spikes 3-6 hours after exercise. Some evidence supports alternate-day dosing to improve fractional absorption, since a dose itself raises hepcidin into the next day.

EA equals energy intake minus exercise energy expenditure, divided by fat-free mass: the energy left over for physiological function after training is paid for. An athlete can hold stable body weight, appearing in energy balance, while running low EA, because the body downregulates processes like reproduction and bone turnover to close the gap.

The cutoff came from short lab studies in a narrow population, and field studies in athletes have often failed to find a clean threshold linking measured EA to hormonal or menstrual disturbance. Free-living EA measurement carries large errors in both intake and expenditure. Current consensus thinking treats LEA as a dose- and duration-dependent exposure rather than a single universal line.

RED-S affects males as well: low energy availability suppresses the hypothalamic-pituitary-gonadal axis in both sexes, hitting testosterone, bone, and performance in men. Experimental work suggests male thresholds may sit lower, around 20-25 kcal/kg FFM/day in high-expenditure sports, but no male-specific threshold is firmly established. The male evidence base lags the female one considerably.

A 2022 systematic review and meta-analysis found acute ketone monoester and precursor ingestion does not improve endurance performance, with most trials null and some negative, including a roughly 2% slower 31km time trial in professional cyclists given a ketone diester. Research interest has shifted toward recovery and overreaching applications, where evidence is preliminary. As a race-day fuel, the claim has not held up.

It raises extracellular bicarbonate and pH, enlarging the gradient between intracellular and extracellular H+. That gradient drives faster lactate and H+ efflux from muscle through monocarboxylate transporters, delaying the intramuscular acidosis that impairs contraction and glycolytic flux. Bicarbonate itself stays extracellular; the buffering benefit is delivered through enhanced export.

Not currently. Meta-analytic data show at most trivial performance differences across phases, with evidence quality rated low and large between-study variation. Substrate use shifts across the cycle are small and readily covered by standard fueling. The supported approach is individual monitoring and cycle awareness, not phase-based macro prescriptions.

Across roughly 1,750 adults, total energy expenditure rose less than activity would predict: energy compensation averaged 28%, meaning around 72% of added activity energy showed up in total expenditure. Compensation appeared partly through reduced basal expenditure and was larger in people with higher adiposity. Added exercise buys fewer net calories than the treadmill display implies.

The model holds that total energy expenditure is regulated within a relatively narrow range: as activity rises chronically, the body trims spending elsewhere (NEAT, inflammation, reproduction, stress axes) rather than adding every activity calorie on top. Practically, exercise-only fat loss plans tend to underdeliver, and dietary intake has to carry most of a deficit. The model is influential but still debated, with some longitudinal data showing little resting-expenditure compensation.

Oral bacteria reduce nitrate to nitrite, which is further reduced to nitric oxide in low-oxygen, low-pH environments such as contracting muscle. Supplementation lowers the oxygen cost of submaximal work by roughly 3-5% and appears to preferentially aid type II fibers. Effects are consistently smaller or absent in elite endurance athletes, whose adaptations may leave little room for improvement.

Gluconeogenesis produces glucose on a grams-per-hour scale suited to maintaining blood glucose, orders of magnitude below the glycolytic flux a hard interval demands from muscle glycogen. Keto-adaptation also downregulates pyruvate dehydrogenase, impairing carbohydrate oxidation when it is needed most. The system defends blood glucose adequately at rest; it cannot bankroll repeated glycolytic efforts.

ATGL performs the first step, triacylglycerol to diacylglycerol, and sets basal lipolytic rate, requiring the coactivator CGI-58. HSL handles the diacylglycerol step and is the catecholamine-sensitive enzyme: beta-adrenergic signalling raises cAMP and PKA, which phosphorylates HSL and perilipin-1 so the droplet coat opens. Insulin opposes the whole system via PDE3B, which degrades cAMP, and adipose lipolysis is suppressed at insulin concentrations below those needed for muscle glucose uptake. Monoacylglycerol lipase completes the third step.

FatMax is the exercise intensity at which whole-body fat oxidation peaks in a graded indirect-calorimetry test, commonly 45-65% VO2max in trained people. Indirect calorimetry reports whole-body substrate use, so it cannot say which tissue or which lipid pool supplied the fat. Individual reliability is poor: a 99-subject day-to-day study reported within-subject coefficients of variation of 21% for peak fat oxidation and 26% for the FatMax intensity, with published estimates ranging roughly 3-26% by protocol. Repeat tests are needed before one number drives training zones.

Muscle mass tracks the integral of synthesis minus breakdown over time, and both terms move. Feeding raises synthesis several-fold and suppresses breakdown modestly; resistance exercise raises both, with synthesis rising further when amino acids are available. Whole-body turnover runs around 3-4 g protein per kg body mass per day, far above intake, because most amino acids are recycled. A trial reporting a synthesis increase says nothing about mass unless breakdown is measured or reasoned about.

Lactate is a continuously produced fuel and signalling molecule, not a dead-end waste product of oxygen shortage. It moves between glycolytic and oxidative fibres and between muscle and heart, liver and brain through monocarboxylate transporters, MCT4 favouring export and MCT1 uptake, then converts back to pyruvate for oxidation. Lactate production rises with glycolytic flux even when oxygen delivery is adequate, so anaerobic threshold language misstates the mechanism. The cell-to-cell shuttle model is mainstream; the intracellular mitochondrial version of it remains debated.

Muscle glycolysis to lactate nets 2 ATP, while hepatic gluconeogenesis from that lactate costs 6 ATP equivalents. Net cost is 4 ATP per glucose, paid by liver oxidative metabolism. The cycle buys fast ATP at the working muscle when glycolytic flux exceeds oxidative capacity, and shifts the energetic bill to a tissue with oxygen to spare. It also prevents lactate accumulation from driving systemic acidosis.

A low insulin to glucagon ratio drives ketogenesis on three fronts at once. Falling insulin releases adipose lipolysis and raises hepatic fatty acid delivery; falling malonyl-CoA removes inhibition of CPT-1 so fatty acids enter the mitochondrion; and mitochondrial HMG-CoA synthase 2, the committed step, is induced and activated under the same conditions. Substrate supply and hormonal signal work together, so neither alone explains the rate.

Muscle glycogen sits in three pools: intramyofibrillar, intermyofibrillar and subsarcolemmal. The intramyofibrillar pool is the smallest, roughly 5-15% of total, but sits beside the triad and correlates with sarcoplasmic reticulum calcium release rate. Ortenblad and Nielsen link depletion of that pool, rather than total glycogen, to reduced tetanic calcium and force loss, which is why fatigue can appear while whole-muscle glycogen still measures adequate.

Variability is large: intra-individual coefficients of variation near 20% and inter-individual near 25% have been reported for white bread under standard methodology, and adding subjects, replicates or longer sampling did not fix it. Values come from a 50 g available-carbohydrate portion of one food eaten alone after an overnight fast, which is not how food is consumed. At group level, carbohydrate amount plus calculated meal GI still explains most of the glycemic response of mixed meals; the failure sits at the individual level.

Colonic fermentation yields acetate, propionate and butyrate in roughly 60:25:15 proportions, feeding colonocytes, supplying gluconeogenic substrate and signalling through free fatty acid receptors. Colonic SCFA infusion in humans raised fasting fat oxidation and resting energy expenditure, and rodent work links butyrate to muscle mass preservation. Performance trials in athletes are scarce, so the athlete-facing claims are extrapolated mechanism rather than demonstrated effect.

Obligatory costs come from digestion, absorption and the ATP spent storing or converting nutrients; a facultative component involves sympathetic activation. Protein runs about 20-30% of its ingested energy, carbohydrate 5-10%, fat 0-3%, and total TEF is around 10% of intake on a mixed diet. Protein is expensive because peptide bond turnover, urea synthesis and gluconeogenesis all consume ATP. Storing dietary fat as body fat is cheap, which is why its thermic cost is near zero.

Falling leptin lowers sympathetic tone and circulating T3 and raises skeletal muscle work efficiency, so expenditure sits below what the new fat mass and fat-free mass predict. Reported magnitudes cluster near 50-150 kcal/day after moderate loss, with the range across studies running roughly 40-230 kcal/day, and much of the early signal shrinking once weight stabilises. Long-term persistence is contested: some cohorts show under 20 kcal/day residual at one to two years, while others report a durable deficit.

Rosenbaum and colleagues held subjects 10% below usual weight and gave leptin to restore pre-loss concentrations. Energy expenditure, skeletal muscle work efficiency, sympathetic tone and thyroid hormone reverted toward pre-loss values. Leptin therefore acts as a signal of energy scarcity when it falls, rather than as a satiety hormone whose excess suppresses appetite. People with obesity are typically hyperleptinemic and resistant, which is why leptin failed as an obesity drug.

Ghrelin comes mainly from the gastric fundus, rises before habitual meals and falls after eating, and is the only established circulating orexigenic gut hormone. Pre-meal rises are partly conditioned by habitual meal timing rather than fixed by an empty stomach, so shifting meal schedules shifts the peaks. After diet-induced weight loss fasting ghrelin rises, and in Sumithran's 62-week follow-up it remained above pre-diet values.

Both are released from intestinal L-cells in response to nutrient delivery. GLP-1 potentiates glucose-stimulated insulin secretion and slows gastric emptying; PYY3-36 acts on hypothalamic Y2 receptors to reduce intake. In Sumithran's 2011 cohort, PYY, CCK, leptin and insulin remained below pre-diet values one year after a ten-week weight loss, while ghrelin and GIP remained raised. Satiety signalling stayed blunted and hunger signalling stayed elevated a full year out.

POMC/CART neurons release alpha-MSH onto MC4R and reduce intake. AgRP/NPY neurons oppose them, with AgRP acting as an inverse agonist at MC4R. Leptin and insulin excite POMC and inhibit AgRP; ghrelin does the reverse. AgRP firing rises with energy deficit and drops on the sight of food before any nutrient is absorbed, which is why appetite anticipates energy state rather than reporting it.

Permissive above a low threshold. Below a minimum concentration synthesis stays low and cannot be raised by amino acids or contraction, while above it adding more insulin does not raise synthesis when amino acids are already delivered. Insulin's separate and dose-related action is suppression of proteolysis, which occurs independently of amino acid availability. The practical read is that adding carbohydrate to an adequate protein dose does not add to the synthetic response.

Through LAT1 (SLC7A5) heterodimerised with CD98 (SLC3A2), an obligatory exchanger that imports large neutral amino acids while exporting glutamine. Muscle-specific SLC7A5 knockout blunts leucine-driven mTOR-S6K signalling, and LAT1 expression rises after feeding and resistance exercise. Because it is an exchanger, import depends on an intracellular pool of exchange substrate as well as on extracellular leucine. Transport capacity therefore gates the signal that plasma leucine concentration alone would predict.

The correlational data run the other way. West and Phillips found acute post-exercise cortisol weakly but positively associated with gains in lean mass and fibre cross-sectional area across a training cohort. Chronic glucocorticoid excess is catabolic through FoxO-driven atrophy gene expression, which is where the gym belief originates, but that is not the transient exercise response. Intramuscular factors such as androgen receptor content predict hypertrophy better than any systemic hormone.

Dose separates the findings. Lilja 2018 gave 1200 mg/day ibuprofen or 75 mg aspirin across eight weeks of knee extensor training: quadriceps volume rose 7.5% on aspirin against 3.7% on ibuprofen with matched training. Later work at over-the-counter doses found limited effect on the hypertrophy mechanisms, and a trial in trained men reported NSAID ingestion augmenting hypertrophy. The defensible position is that chronic high-dose use around training is unwise, while occasional standard doses are not shown to be harmful.

Smith's 2011 trials found eight weeks of 4 g/day omega-3 left basal synthesis unchanged but augmented the response to a hyperinsulinaemic-hyperaminoacidaemic clamp, with greater mTORC1 signalling. Replication has been inconsistent: one trial in young men found fish oil altered anabolic signalling after resistance exercise and feeding without changing myofibrillar synthesis. Meta-analytic work reports heterogeneity in population, dose and outcome, so the sensitising effect is plausible but not settled.

Controlled trials in healthy people do not show it. A systematic review of 15 RCTs found no significant effect of added linoleic acid on CRP, fibrinogen, PAI-1, cytokines, soluble adhesion molecules or TNF-alpha. Arachidonic acid supplementation raises tissue arachidonic acid without raising inflammatory cytokine expression. The linoleic to arachidonic to eicosanoid pathway is real, but human conversion is low and the downstream inflammatory claim is not supported by intervention data.

Contested on the detection side. Many early immunohistochemistry results used antibodies later shown to bind proteins other than VDR, and VDR mRNA sits near the detection limit in human muscle biopsies. Muscle-specific knockout and overexpression work in mice supports a functional receptor present at low abundance. Supplementation trials show reduced falls in deficient or institutionalised older adults, while performance benefit in vitamin D replete athletes is not demonstrated.

Dermal calcium loss in sweat plus intestinal absorption that lags demand can lower ionised calcium during long sessions, and PTH rises to defend it by mobilising skeletal calcium. Trials giving calcium before rather than after prolonged exercise blunt both the PTH rise and the bone resorption marker response. Mechanical loading itself is osteogenic, so the practical target is impact or heavy loading plus calcium and vitamin D sufficiency, with calcium delivered before long sessions rather than only after.

Folate carries one-carbon units for purine and thymidylate synthesis; B12 is the cofactor for methionine synthase, which remethylates homocysteine to methionine and regenerates tetrahydrofolate. Without B12 the folate pool is trapped as 5-methyl-THF and cannot be recycled, the methyl-folate trap, so homocysteine rises. Generous folate intake can correct the macrocytic anaemia while B12-driven neurological damage continues, which is why B12 status is measured rather than inferred from blood counts.

Zinc is required by hundreds of enzymes and for thymulin activity and lymphocyte function, and frank deficiency impairs immunity. Zinc acetate or gluconate lozenges at high daily doses, started within about 24 hours, shortened common cold duration in meta-analyses of general populations, an effect attributed to local ionic zinc in the pharynx rather than to systemic repletion; the 2024 Cochrane review rates the evidence low certainty. Routine high-dose supplementation in zinc-replete athletes has no demonstrated performance or infection benefit, and sustained intakes above roughly 40 mg/day impair copper status.

Magnesium is the counter-ion of ATP, since Mg-ATP is the actual enzyme substrate, a cofactor for over 300 enzymes across glycolysis and oxidative phosphorylation, and a modulator of NMDA and calcium channels. Correcting genuine deficiency improves exercise performance, and sweat and urinary losses may raise athlete requirements by 10-20% with many athletes under intake targets. In replete regular exercisers supplementation is not ergogenic, and one crossover trial reported modest negative effects on cycle performance and mitochondrial respiration.

Sweat sodium concentration falls, often within a few days of repeated heat exposure. The secretory coil produces near-isotonic fluid and sodium is reabsorbed along the duct through luminal ENaC and basolateral Na/K-ATPase, with aldosterone raising ENaC expression and membrane trafficking. Acclimated athletes conserve sodium at higher sweat rates through greater duct responsiveness. Individual sweat sodium spans a wide range, so population averages make poor prescriptions.

Fluid intake beyond thirst combined with non-osmotic arginine vasopressin release, which retains free water and satisfies the diagnostic criteria for SIADH. Sweat sodium losses contribute but are rarely sufficient alone; the dilution comes from water retained and consumed. Muscle-derived IL-6 is one proposed non-osmotic trigger for AVP release. Sodium supplementation does not reliably prevent the condition when drinking remains excessive.

Symptomatic hyponatremia has caused documented race deaths from cerebral oedema; deaths attributed to exercise-associated dehydration in the same events are not established. Dehydration around 2-3% of body mass costs performance under some conditions and raises thermal strain, but it self-corrects after the event, while hyponatremic encephalopathy can kill within hours. That asymmetry, not a claim that dehydration stopped mattering, is why drink-to-thirst replaced the 1990s advice to drink as much as tolerable.

Early trials from a small number of groups reported large gains in lean mass and strength, several of implausible magnitude in already-trained subjects. Independent replication in trained athletes largely failed, and current systematic reviews report small and inconsistent effects with the clearest signal in clinical and older populations rather than in athletes. Umbrella reviews of those meta-analyses flag a substantial share of low or critically low quality evidence. The IOC does not group HMB with the strongly supported ergogenic aids.

Leucine triggers mTORC1 signalling, but synthesis requires all the amino acids in the sequence; a BCAA-only dose signals for construction without supplying materials, so the remainder has to come from breakdown of existing protein. Jackman 2017 found BCAAs after resistance exercise raised myofibrillar synthesis 22% over placebo, roughly half the response reported for a whey dose containing comparable BCAA content. For anyone already eating adequate protein there is no case for BCAAs over whole protein or complete EAAs.

Glutamine fuels enterocytes and lymphocytes and plasma concentrations fall under heavy training loads, which generated the immune-support hypothesis in the 1990s. A 2018 systematic review and meta-analysis found no effect on immune markers, aerobic performance or body composition, and the IOC groups glutamine with insufficiently evidenced supplements. Isolated trials report benefits for mucosal immunity or gut permeability in specific populations, so the fair summary is unproven rather than refuted.

Arginine undergoes heavy first-pass metabolism in gut and liver, including by arginase, so systemic delivery is limited and the high doses needed cause GI distress. Citrulline escapes presystemic breakdown, is absorbed well and reabsorbed by the kidney, where it is converted to arginine. Citrulline supplementation therefore produces higher steady-state plasma arginine than an equimolar arginine dose. Whether that translates into performance benefit is a separate question, and the effect sizes there are modest.

Meta-analyses report faster recovery of muscle strength and lower reported soreness after strenuous exercise, with the largest effects when juice is taken for several days before the bout rather than only after it. GRADE assessments rate certainty low to moderate on heterogeneity and imprecision, and most trials are small. The proposed anthocyanin antioxidant mechanism raises the same adaptation-blunting question as antioxidant megadoses and cold water immersion during adaptation phases.

Curcumin is poorly water soluble, unstable at intestinal pH, and heavily conjugated by glucuronidation and sulfation in gut wall and liver, so unconjugated plasma concentrations remain in the low nanomolar range after gram doses. The widely quoted 2000% increase from 20 mg piperine traces to a single small 1998 study, and piperine works by inhibiting glucuronidation, which also affects clearance of other drugs. Recent pharmacokinetic reappraisals found free curcumin still around 100-fold below the concentrations used in the cell studies, including with enhanced formulations.

Ethanol has no storage pool and no meaningful excretion route, so it is oxidised on arrival, largely to acetate released into plasma. In Siler's tracer work, 24 g alcohol raised plasma acetate flux to 2.5 times baseline, cut adipose non-esterified fatty acid release by 53% and reduced whole-body lipid oxidation by 73%. De novo lipogenesis from ethanol was modest, so the fat-gain route is displaced oxidation of dietary fat rather than conversion of alcohol into body fat.

In people with fatty liver, isotope studies attribute roughly a quarter of hepatic triglyceride to de novo lipogenesis, with most of the remainder from adipose free fatty acids and dietary fat. Fructose stimulates hepatic DNL more than glucose and does so dose-dependently, yet the fraction of an ordinary fructose load converted to fat is small, and tracer work shows the lipogenic effect is not explained by substrate provision alone. Isocaloric substitution trials show far smaller effects than hypercaloric fructose overfeeding, so dose and energy balance carry the claim rather than the sugar's identity.

It gave 120 sweetener-naive adults saccharin, sucralose, aspartame or stevia below the acceptable daily intake for two weeks. All four shifted gut microbiome composition, but only saccharin and sucralose impaired glycemic responses, and microbiota transfer from top and bottom responders into germ-free mice reproduced the donors' glycemic patterns. Responses were highly individual, and the trial was short with a small per-arm sample, so it demonstrates a causal mechanism rather than a population health outcome.

Within days. Five-day controlled protocols at around 15 kcal per kg fat-free mass per day lower P1NP, a formation marker, and raise CTX, a resorption marker, shifting turnover toward resorption long before any density change is measurable, with similar shifts reported in men in later low energy availability trials. Ihle and Loucks showed a dose-response across energy availability levels in exercising women using PICP and NTX, with formation suppressed at milder restriction than that which raised resorption. Markers move far faster than DXA, but they are noisy and require standardised fasted timing to interpret.

At matched relative intensity, women typically show a lower respiratory exchange ratio: more fat and less carbohydrate oxidised, with lower muscle glycogen use and lower hepatic glucose output. Higher intramyocellular lipid content, greater adipose lipolysis and higher beta-oxidation enzyme protein such as VLCAD and MCAD accompany the difference, with 17-beta-estradiol implicated largely from rodent work. The effect is modest and not universal: some studies prescribing intensity by heart rate or ventilatory threshold find no sex difference, and matching training status and habitual diet narrows it.

In a crossover trial each participant receives both treatment and placebo, so each serves as their own control, removing between-subject variance and raising statistical power at small sample sizes. The cost is that crossovers require an adequate washout period so the first treatment does not carry into the second, and counterbalanced treatment order so learning or training effects do not masquerade as treatment effects. Substances with long tissue retention, creatine being the standard example, need washouts of a month or more, which makes crossover designs impractical for them.

FFQs rely on memory and portion estimation over weeks to months, producing large random and systematic error in nutrient estimates. Energy intake is systematically underreported, and the underreporting tends to worsen with higher body weight, which biases diet-disease associations. Because the error is correlated with characteristics of interest rather than purely random, calibration against recovery biomarkers such as doubly labeled water or urinary nitrogen shows FFQs can misestimate intake substantially at the individual level.

Supplement users tend to exercise more, smoke less, earn more, and see doctors more often than non-users, so their better outcomes reflect the whole lifestyle package rather than the pill. Statistical adjustment reduces this but never removes it: covariates are measured with error and unmeasured factors remain, which is residual confounding. This is a main reason observational benefits of supplements such as vitamin E and multivitamins failed to reproduce in randomized trials.

A p-value only says the result is unlikely under the null; with enough participants a trivial difference of a fraction of a percent reaches significance. What matters for sport is the effect size with its confidence interval, judged against the smallest worthwhile change, which in elite endurance events can be under 1%. A wide confidence interval spanning trivial to large effects means the trial was uninformative regardless of the p-value.

Trials with positive results are more likely to be submitted, published, and cited than null trials, so the published record overstates true effects. Small studies with null or negative findings go missing, which shows up as asymmetry in a funnel plot of effect size against study precision. These small-study effects mean pooled estimates in supplement meta-analyses often shrink when analysis is restricted to larger, pre-registered trials.

Reviewing 206 articles on soft drinks, juice, and milk, they found studies funded entirely by industry had higher odds of conclusions favorable to the sponsor than studies with no industry funding: an odds ratio of 4.37 unadjusted, rising to 7.61 after accounting for beverage type, publication year, and author conflicts of interest. Among interventional studies, none of the all-industry-funded papers reached an unfavorable conclusion. The bias showed up in the framing of conclusions rather than in fabricated data, so study questions, comparators, and interpretation can all tilt favorable.

I-squared estimates the proportion of variation across studies due to true between-study differences rather than chance. High values mean the trials differ in dose, population, protocol, or quality, so a single pooled number can be an average of incomparable things. The garbage-in problem is that pooling many weak, biased trials produces a precise-looking but unreliable estimate; meta-analysis cannot repair flaws in the underlying studies.

A 2020 systematic review pooling 32 studies and about 1,500 participants found small-to-moderate effects for both placebo (d around 0.36) and nocebo (d around 0.37) conditions, with nutritional ergogenic aids at roughly d 0.35. That magnitude overlaps with the measured effect of several legal supplements. Belief about what was ingested can shift pain perception and pacing, which is why deceptive and balanced-placebo designs exist.

Caffeine produces perceptible arousal, alertness, and sometimes jitters, so habitual users often identify their condition, and correct guesses add expectancy to the measured effect. Sodium bicarbonate causes dose-dependent gastrointestinal symptoms such as bloating and diarrhea that a placebo rarely mimics, unblinding participants in the other direction. Few trials formally test blinding success, so published effect sizes for both likely carry an expectancy component.

Elite athletes are too few and too heterogeneous for adequately powered group trials, and group means can hide meaningful individual responses. Single-subject designs use repeated baseline and intervention phases within one athlete, with multiple crossovers, to estimate that athlete's own response against their own variability. The trade-off is limited generalizability and vulnerability to time-varying confounders such as training phase.

Indirect calorimetry infers energy expenditure and fuel mix from VO2 and VCO2, assuming a physiological steady state and that expired CO2 reflects substrate oxidation. During heavy exercise, bicarbonate buffering of hydrogen ions releases extra non-metabolic CO2, pushing RER above 1.0 and invalidating substrate calculations. Protein oxidation is also usually ignored, so tables report nonprotein RER, a small but real simplification.

DLW tracks the differential elimination of stable hydrogen and oxygen isotopes to estimate CO2 production, giving free-living total energy expenditure with roughly 1-3% accuracy and 2-8% precision over 1-2 weeks. It is the reference method for validating intake and wearable estimates. Its limits: it yields only a total, with no breakdown into activity, exercise, or resting components, no day-by-day resolution, and it is expensive.

DXA soft-tissue models assume fat-free mass hydration is constant near 73%, but real hydration ranges from about 67% to 85%. Acute water ingestion of 500 ml before a scan significantly inflates fat-free mass, and controlled feeding studies raised lean soft tissue estimates by up to about 1.7% overall (higher in some regions and individuals) while lowering fat mass estimates by up to about 3%. Values returned to baseline after an overnight fast, so standardized conditions, fasted, rested, and consistent time of day, restore comparability between scans.

BIA measures the impedance of a small current and predicts total body water, then fat-free mass, through population-derived regression equations. Hydration status, recent food and fluid, exercise, skin temperature, and electrode or stance position all shift the reading, and the equations add error when the athlete differs from the reference population. Group averages can be acceptable while individual error spans several percentage points of body fat.

Calipers measure double folds of subcutaneous fat at defined sites; converting the sum to percent body fat requires regression equations that assume fixed relationships between subcutaneous and internal fat and population-typical fat patterning. Those conversions add error, so many practitioners track the raw sum of sites in millimeters over time instead. With a trained, consistent tester, the sum of skinfolds is a sensitive change measure even when the derived percentage is questionable.

A 2023 systematic review with meta-analysis in athletes found predicted values from the Cunningham equation, which predicts from fat-free mass, did not differ significantly from measured RMR, while Mifflin-St Jeor significantly underestimated it. Fat-free mass is the dominant driver of resting metabolism, so equations built on total body weight in general populations tend to underpredict in muscular athletes. Individual studies disagree in some sub-populations, and even the best equation can miss an individual by hundreds of kcal, so measurement wins when stakes are high.

Shcherbina and colleagues (2017) tested seven wrist devices against lab references in 60 diverse volunteers: six of the seven kept median heart rate error under 5%, but no device kept energy expenditure error under 20%. Median calorie error ran from about 27% for the best device to about 93% for the worst, and error was highest during low-intensity tasks. Optical HR sensing is a mature measurement; calorie estimates are model outputs stacked on top of it.

US regulation tolerates energy content up to 20% above the stated label value. Analytical work found frozen supermarket meals averaged about 8% more calories than labeled, while restaurant items were worse: measured energy averaged about 18% above stated values, with individual dishes off by hundreds of kcal. Database entries are averages over brands, seasons, and preparations, so an individual portion can differ meaningfully from the logged number.

Kjeldahl-type methods measure nitrogen, and multiplying by 6.25 assumes all proteins are 16% nitrogen and that all measured nitrogen comes from protein. Neither holds: amino acid profiles vary, so food-specific factors run well below 6.25 for many foods, and foods contain non-protein nitrogen from ammonia, nucleotides, and other compounds. The fixed factor therefore overestimates protein in most foods, which is why some analysts argue for food-specific factors or a lower default near 5.6.

NCAA wrestling certifies a minimum competition weight corresponding to 5% body fat for men and 12% for women, paired with a urine specific gravity check (1.020 or lower) to block dehydrated certification, following weight-cutting deaths in 1997. Ski jumping ties maximum ski length to body mass index: from 2004 the FIS set maximum ski length at 145% of height for jumpers at or above a minimum BMI, with shorter skis below it, removing the aerodynamic payoff of extreme leanness. Both make the underweight strategy unprofitable rather than merely discouraged.

USG lags behind plasma osmolality during acute dehydration and rehydration, so a spot sample can misrepresent current body water. First-morning samples run concentrated because of overnight fluid abstinence, inflating apparent dehydration, and larger athletes tend toward higher values regardless of hydration. Validation work classifying athletes at standard cutoffs such as 1.015 and 1.020 got only about 65% correct.

Exercise and dehydration contract plasma volume as water moves out of the vasculature, concentrating everything dissolved in the remaining plasma. Marker concentrations can rise 10% or more with no change in total content, faking increases in electrolytes, proteins, and hormones. The Dill and Costill (1974) equation uses pre and post hemoglobin and hematocrit to estimate the volume shift and correct concentrations, though it suits plasma and serum markers better than whole-blood measures such as lactate and white cells.

Ferritin is an acute-phase reactant that rises with inflammation, so hard training and racing raise it independent of iron stores. Depending on duration and intensity, ferritin may stay flat, rise by roughly a quarter and settle within a day, or roughly double after ultramarathon-length efforts and take several days to return to baseline. A sample drawn in that window can mask true iron deficiency, so best practice is testing after rest days, at a consistent morning time, ideally alongside an inflammation marker such as CRP to flag confounded results.

External quality schemes and method-comparison studies show automated immunoassays can deviate from LC-MS/MS reference methods by roughly 15% in either direction, with individual platforms documented well outside that band. Assays also struggle at range extremes and with the D2 form. An athlete near a clinical threshold can be classified deficient on one platform and sufficient on another, so trends should be followed on the same assay and borderline values not overinterpreted.

Cowley and colleagues (2021) audited over 5,200 publications and 12.5 million participants across six major journals from 2014-2020: 34% of participants were female, and only 6% of studies were female-only versus 31% male-only. Sports nutrition guidelines built on this literature therefore default to male-derived data. The consequence is real uncertainty about whether dosing and fueling recommendations transfer, not evidence that they fail.

Calendar counting misclassifies phase because cycle and phase lengths vary and anovulatory cycles occur without symptoms; best-practice methodology requires cycle tracking, urinary LH ovulation confirmation, and serum estradiol and progesterone verification. Most older studies used calendar counting alone, undermining phase-based conclusions. Hormonal contraceptive users have a different hormonal profile, suppressed endogenous cycling plus exogenous hormones, so pooling them with naturally cycling participants blends two distinct physiologies.

Associations between gut microbes and fitness, such as Veillonella enrichment in marathon runners, are real, but the mechanistic follow-up was done in mice and human performance trials remain few, small, and inconsistent. Differences in sequencing methods and analysis pipelines, plus large interindividual variability, hamper replication. Probiotics have moderate evidence for reducing illness burden in athletes, which can protect training, while 'optimize your microbiome for performance' products outrun the trial evidence.

Most DTC panels type a handful of common variants, FTO and CYP1A2 among them, and issue diet advice from single-SNP associations that explain a small fraction of trait variance. Interventional evidence that genotype-matched diets outperform standard personalized advice is lacking, and reviews of the sector report companies selling recommendations with little supporting literature. Some single findings, CYP1A2 and caffeine response included, remain contested rather than settled.

Food4Me (1,607 adults, seven countries) found personalized advice improved dietary behavior versus generic advice, but adding phenotype or genotype layers gave no extra benefit, and weight loss did not differ across arms at six months. PREDICT documented large, repeatable interindividual variation in postprandial responses to identical meals, on the order of 100% for triglycerides and 68% for glucose, with meal macronutrients explaining only part of the variance. Neither established that commercial personalized scores improve athletic performance or long-term health outcomes.

CGMs read interstitial, not blood, glucose, with lag times reported at 5-25 minutes that widen during exercise as blood flow, temperature, and pH shift. Reviews conclude there is insufficient evidence that CGM-guided interventions improve performance or recovery in healthy athletes, and CGM does not measure muscle glycogen or carbohydrate flux, which limits it as a fuel sensor. Interpretation frameworks for healthy athletes do not yet exist, so the devices currently generate data without decisions.

Capillary blood meters measure beta-hydroxybutyrate and are the practical reference; nutritional ketosis is conventionally defined from about 0.5 mmol/L. Urine strips detect acetoacetate and lose validity with keto-adaptation as more acetoacetate is converted to BHB and reabsorbed, and hydration dilutes readings. Breath acetone correlates with ketosis but is semi-quantitative and sensitive to measurement conditions, so it suits trend-watching rather than thresholds.

Lactate distributes unevenly between plasma and red cells, so plasma values run substantially higher than whole-blood values, and the size of the gap depends on hematocrit. Analyzers differ in what they measure and how they express it, some are hematocrit-sensitive and some are not, and handheld devices vary in precision across intensities. Threshold testing is therefore only internally consistent: same device, same sample site, same protocol, every test.

Systematic review evidence shows MCT supplementation raises blood ketones but does not improve endurance performance, and most trials found no effect on respiratory exchange ratio, substrate oxidation, or lactate. The induced ketones are not used as a primary fuel during acute exercise, and glycogen sparing has not been demonstrated. Gastrointestinal distress is dose-dependent, with around 30 g treated as the practical ceiling, which capped doses in the research itself.

Ribose feeds nucleotide synthesis toward adenine nucleotide resynthesis, and intense training does transiently lower muscle ATP. One trial found faster ATP restoration with ribose after intense intermittent training, yet mean and peak power output at 72 hours matched placebo, and other trials found no effect on ATP recovery or repeated maximal exercise. The post-training ATP dip does not limit high-intensity performance, so restoring it faster buys nothing.

Oral and sublingual IGF-1 has negligible systemic absorption, no controlled evidence supports performance benefits, and laboratory analysis of a prominent product found no deer-sequence IGF-1 at all. WADA has urged athletes to be extremely vigilant with these products because IGF-1 is prohibited and because contamination with undeclared substances is a real manufacturing hazard. The category combines an ineffective delivery route with a doping risk.

These programs test product batches for a panel of prohibited substances and audit manufacturing, which meaningfully lowers, but does not eliminate, contamination risk. They do not verify efficacy, do not test every unit produced, and cannot cover every substance on the prohibited list. Under strict liability the athlete remains responsible for any positive test, so certification is risk management, not immunity.

Geyer and colleagues analyzed 634 non-hormonal supplements bought in 13 countries between late 2000 and late 2001: 94 of them, 14.8%, contained anabolic androgenic steroids not declared on the label. Products from companies also selling prohormones were dirtier (21.1%) than those from companies that did not (9.6%), pointing at cross-contamination in shared manufacturing. This study is the standard citation for why unscreened supplements are a doping hazard.

Caffeine was prohibited in competition above a urinary threshold through the 1980s and 1990s, with the threshold set at 12 micrograms per milliliter in 1987, until WADA removed it from the Prohibited List effective 1 January 2004. It sits on the Monitoring Program, where urinary concentrations are still tracked in case usage patterns warrant re-listing. Performance doses of 3-6 mg/kg sit well below the old threshold anyway.

Pooled dose-response analysis by Moore and colleagues placed the per-meal breakpoint for maximal MPS at about 0.24 g/kg body mass in younger men versus about 0.40 g/kg in older men, the anabolic resistance of aging expressed as a dose shift. Consistent with that, the Morton meta-regression found protein supplementation's effect on fat-free mass gains shrank with increasing age. Older athletes therefore need larger boluses per meal, not merely more daily total.

The Devries meta-analysis of 28 trials in healthy adults found the change in glomerular filtration rate did not differ between higher-protein diets (1.5 g/kg or more) and lower or normal-protein diets; the transient GFR rise with protein is adaptive hyperfiltration, not injury. The broader lesson generalizes: surrogate signals, a raised GFR, an MPS spike, a lactate shift, are not outcomes. Sports nutrition claims stand or fall on hard endpoints, kidney disease incidence or race-day performance, and the surrogate-to-outcome jump is where most claims quietly fail.