Skip to content
The Longevity Archive
browse 192

The biology of fasting

The growth hormone response

Also called: GH, Somatotropin, Counter-regulatory response

Written by Reviewed Sep 2026

Growth hormone secretion rises several-fold within two days of a fast, and the growth factor it normally works through does not follow it, which is the part that changes what the rise means.

What it is

Blood glucose is defended asymmetrically. One hormone lowers it, insulin. Several raise it: glucagon, adrenaline and noradrenaline, cortisol, and growth hormone. Those are the counter-regulatory hormones, and a fast engages all of them, on different timescales and with different jobs.

Growth hormone is released in pulses from the anterior pituitary, under opposing control by hypothalamic GHRH, which stimulates, and somatostatin, which inhibits, with ghrelin as a further stimulatory input. Because it is pulsatile, a single blood sample is close to meaningless. The measurements worth citing come from sampling every few minutes for 24 hours and then reconstructing the underlying secretion mathematically, and both landmark human studies did exactly that.

Growth hormone normally works largely by instructing the liver to make IGF-1, insulin-like growth factor 1, which is what carries most of its anabolic effect to tissue. During a fast that link is deliberately broken.

The machinery

Fasting raises growth hormone by changing the shape of its secretion, not by slowing its removal. Deconvolution analysis of five-minute sampling over 24 hours on the second day of a fast in nine normal men showed the 24 hour production rate rising roughly five-fold, driven by two separate changes: the number of secretory bursts per 24 hours more than doubled, and the mass of hormone released per burst also roughly doubled through higher burst amplitude with unchanged burst duration. The gaps of quiescence between volleys shortened. The half-life of growth hormone in blood was unchanged, which rules out the alternative explanation that hormone simply accumulates because disposal slows. The proposed mechanism is more frequent GHRH release plus longer and deeper periods of somatostatin withdrawal.

What growth hormone then does during a fast is mobilise fat. It is lipolytic and it antagonises insulin's action on glucose. What it does not do during a fast is drive IGF-1, because the liver becomes growth hormone resistant, through downregulation of the growth hormone receptor and upregulation of IGF binding protein 1. The logic is coherent: the lipolytic and glucose-sparing arm is useful to a fasting organism, and the growth-promoting arm is exactly what a starving organism should not be spending resources on. The system separates the two.

The protein-sparing arm works through fat, not through muscle, and that has been tested directly by blocking it. In 7 normal subjects through a 37 hour fast, blocking lipolysis with acipimox raised urea excretion, serum urea and muscle protein breakdown by almost 50%. Adding growth hormone while lipolysis was blocked did nothing to indices of protein degradation. Restoring free fatty acids with intralipid brought whole-body protein degradation back down by 10 to 15%. The authors' conclusion is that blocking lipolysis eliminates the ability of growth hormone to restrict fasting protein loss, and that stimulating lipolysis is its principal protein-conserving mechanism.

What switches it on

Falling glucose and falling insulin, acting through the hypothalamus. The whole counter-regulatory set is engaged in sequence: glucagon first and fastest, then catecholamines and cortisol on a circadian-modulated pattern, with growth hormone rising into the second day.

On the timing
This is one of the few mechanisms on the site with a genuine human clock, because 24 hour sampling studies exist at 24 hours, 48 hours and 5 days. A 24 hour water-only fast significantly increased growth hormone in a randomised crossover of 30 healthy adults, although that paper prints only the p value and not the magnitude, so no percentage from it is quoted here. By the second day the 24 hour production rate is roughly five-fold higher. Across a 72 hour fast in 33 subjects, growth hormone secretion increased from the first day to the third but decreased from the second to the third, so the response is not monotonic. IGF-1 is unchanged at 56 hours and has fallen roughly 41% by day 5.

What has been measured

In people

This is among the best characterised areas of human fasting physiology, because it was worked out with catheterisation and frequent sampling in metabolic wards between the 1960s and the 1990s. Every quantitative statement below is a measurement in people.

  • Two day fast, 9 normal men, serum sampled every 5 minutes for 24 hours on a fed control day and on day two, analysed by multi-parameter deconvolution. The 24 hour endogenous production rate rose from 78 plus or minus 12 to 371 plus or minus 57 micrograms per litre of distribution volume, a five-fold increase, P = 0.0001. Secretory bursts per 24 hours rose from 14 plus or minus 2.3 to 32 plus or minus 2.4, P = 0.0006. Mass per burst rose from 6.3 plus or minus 1.2 to 11 plus or minus 1.6, P = 0.002. The interval between secretory volleys shortened from 143 plus or minus 14 to 88 plus or minus 4.2 minutes. Half-life was unchanged at 18 plus or minus 2.2 versus 20 plus or minus 1.5 minutes, P = 0.47. Serum IGF-1 after 56 hours of fasting was unchanged (Journal of Clinical Endocrinology and Metabolism, 1992).
  • Five day fast, 6 normal adult men, 24 hour sampling analysed by both discrete pulse detection and Fourier time series. Pulse frequency rose from 5.8 plus or minus 0.7 to 9.9 plus or minus 0.7 pulses per 24 hours, P = 0.028. Integrated 24 hour concentration rose from 2.82 plus or minus 0.50 to 8.75 plus or minus 0.82 micrograms times minute per mL, P = 0.0002. Maximal pulse amplitude rose from 5.9 plus or minus 1.1 to 12.3 plus or minus 1.6 ng/mL, P < 0.005. Over the same days, somatomedin C, which is IGF-1, fell from 1.31 plus or minus 0.22 to 0.77 plus or minus 0.18 U/mL, glucose fell from 4.9 to 3.2 mmol/L, and free fatty acids rose from 0.43 to 1.55 mmol/L (Journal of Clinical Investigation, 1988).
  • 72 hour fast, 33 healthy subjects, 16 male and 17 female, with reference intervals for the whole counter-regulatory set. Glucose, insulin, C-peptide and proinsulin all fell (P < 0.001). Glucagon and free fatty acids rose linearly. Adrenaline, noradrenaline and cortisol rose with a clear underlying circadian rhythm superimposed, which means a single timed sample can badly mislead. Growth hormone secretion increased from the first to the third day (P < 0.05) but decreased from the second to the third (P = 0.03). Males had higher glucose and glucagon and lower free fatty acids during the fast (American Journal of Physiology: Endocrinology and Metabolism, 2001).
  • The blocking experiment, 7 normal subjects, 37 hour fast, four conditions. Blocking lipolysis with acipimox raised urea excretion, serum urea and muscle protein breakdown by almost 50%. Adding growth hormone during fasting with antilipolysis did not influence indices of protein degradation. Restoring high free fatty acid levels with intralipid lowered urea and cut whole-body protein degradation by 10 to 15%. Stimulation of lipolysis is the principal protein-conserving mechanism of growth hormone (Journal of Clinical Endocrinology and Metabolism, 2003).
  • The complementary experiment, 8 normal subjects, 40 hour fast with growth hormone suppressed by somatostatin and then replaced, with insulin and glucagon replaced identically in both arms so only growth hormone differed. Urinary urea excretion was 392 plus or minus 44 mmol per 24 hours at baseline, 440 plus or minus 32 fasted, 609 plus or minus 76 fasted without growth hormone, and 408 plus or minus 36 fasted with growth hormone replaced (P < 0.05). Suppressing growth hormone during a 40 hour fast raised urea-nitrogen excretion by roughly 50%, so growth hormone is a genuine component of physiological protein conservation (Diabetes, 2001).
  • What happens when the system is overridden. Human growth hormone administered to obese subjects during 5 to 6 weeks of starvation produced a 2 to 3 fold rise in insulin preceding a 50% rise in blood glucose, persistent lipolysis, and blood beta-hydroxybutyrate and acetoacetate rising 20 to 40% to total blood ketoacid concentrations of 10 to 12 mmol/L with ketonuria of 150 to 320 mmol per day. Subjects reported nausea, vomiting, weakness and myalgias. Despite a 50% reduction in urea excretion, total nitrogen loss was unchanged because urinary ammonia rose 50% (Journal of Clinical Investigation, 1971).
  • 24 hours, randomised crossover, 30 apparently healthy volunteers, one day of water-only fasting against a day of usual eating. Growth hormone rose significantly (p = 1.1 x 10 to the minus 4). Haemoglobin, red cell count, haematocrit, total cholesterol and HDL all rose, while triglycerides, bicarbonate and weight fell. The rise in haemoglobin, red cell count and haematocrit together is the signature of plasma volume contraction rather than red cell production. The magnitude of the growth hormone change is not given in the abstract and is therefore not quoted here (Nutrition, Metabolism and Cardiovascular Diseases, 2013).

In other species and in cell culture

The human data here are good enough that this page does not lean on other species. One species note is load bearing: the pulsatile architecture of growth hormone secretion differs substantially between rat and human, which is stated as the motivation for running the 5 day study in people in the first place.

  • Rat and human growth hormone secretion have different pulsatile architectures, which is why the human 24 hour sampling studies were designed as they were (Journal of Clinical Investigation, 1988).
  • Growth hormone receptor regulation and hypothalamic control of pulsatility have an extensive supporting literature in rodents and cell systems. No specific identifiers from that literature are cited here, because the human measurements are direct and sufficient.

Why it matters

The claim that fasting raises growth hormone is true, and the peer-reviewed numbers behind it are large. A roughly five-fold rise in 24 hour secretion by the second day, with clearance unchanged, is a real and substantial physiological event.

What it means is the part that gets lost. Growth hormone normally acts through hepatic IGF-1, and during a fast IGF-1 does not move at 56 hours and falls by day 5. The liver has become growth hormone resistant on purpose. So the surge is not an anabolic state. It is a fat-mobilising, glucose-sparing state that has been decoupled from growth.

And the protein-sparing effect, which is the reason people care, works by liberating fatty acids rather than by acting on muscle. Block lipolysis and the effect disappears entirely, which is about as direct a test as human physiology allows.

A claim you will see repeated

The figures usually quoted are that growth hormone rises 1,300% in women and 2,000% in men. Those numbers are real and they are quoted accurately, but they come from a press release issued by the Intermountain Medical Center Heart Institute in April 2011, describing work presented at a cardiology conference. Three things travel badly. The percentages are for a 24 hour fast, and the claim is almost always attached to multi-day fasting instead. It is a conference press release rather than a peer-reviewed paper a reader can check. And the same release reports that LDL cholesterol rose 14% and HDL rose 6% during the fast, a detail that never travels with the growth hormone number. The researchers' own closing line was that it is not time to start a fasting diet just yet. The checkable figure is the five-fold rise in 24 hour production rate on day two, from nine men sampled every five minutes. The second misconception is mechanistic. The fasting growth hormone surge is widely said to protect muscle directly. The blocking experiment shows otherwise: with lipolysis blocked, adding growth hormone did nothing at all to protein degradation. It conserves protein by mobilising fat.

Citations

  1. Human1992
    Augmented growth hormone (GH) secretory burst frequency and amplitude mediate enhanced GH secretion during a two-day fast in normal men

    Journal of Clinical Endocrinology and Metabolism

    9 normal men, serum GH every 5 minutes for 24 hours on a fed day and on the second day of a fast, multi-parameter deconvolution. 24 hour production rate 78 plus or minus 12 to 371 plus or minus 57 micrograms per litre of distribution volume, a five-fold rise, P = 0.0001. Bursts per 24 hours 14 plus or minus 2.3 to 32 plus or minus 2.4, P = 0.0006. Mass per burst 6.3 plus or minus 1.2 to 11 plus or minus 1.6, P = 0.002. Half-life unchanged, 18 plus or minus 2.2 versus 20 plus or minus 1.5 minutes, P = 0.47. Serum IGF-1 unchanged after 56 hours.

  2. Human1988
    Fasting enhances growth hormone secretion and amplifies the complex rhythms of growth hormone secretion in man

    Journal of Clinical Investigation

    6 normal adult men, 24 hour sampling on a control fed day and on days 1 and 5 of a 5 day fast, analysed by discrete pulse detection and Fourier time series. Pulse frequency 5.8 plus or minus 0.7 to 9.9 plus or minus 0.7 per 24 hours, P = 0.028. Integrated 24 hour concentration 2.82 plus or minus 0.50 to 8.75 plus or minus 0.82 micrograms times minute per mL, P = 0.0002. Maximal pulse amplitude 5.9 plus or minus 1.1 to 12.3 plus or minus 1.6 ng/mL, P < 0.005. Somatomedin C fell 1.31 plus or minus 0.22 to 0.77 plus or minus 0.18 U/mL, glucose 4.9 to 3.2 mmol/L, free fatty acids 0.43 to 1.55 mmol/L.

  3. Human2001
    Reference intervals for glucose, beta-cell polypeptides, and counterregulatory factors during prolonged fasting

    American Journal of Physiology: Endocrinology and Metabolism

    33 healthy subjects, 16 male and 17 female, across a 72 hour fast. Glucose, insulin, C-peptide and proinsulin all decreased (P < 0.001). Glucagon and free fatty acids increased linearly. Adrenaline, noradrenaline and cortisol increased with a clear circadian rhythm superimposed. Growth hormone secretion increased from the first to the third day (P < 0.05) but decreased from the second to the third (P = 0.03). Males had higher glucose and glucagon and lower free fatty acids (P < 0.05); higher BMI produced higher insulin and C-peptide.

  4. Human2003
    The decisive role of free fatty acids for protein conservation during fasting in humans with and without growth hormone

    Journal of Clinical Endocrinology and Metabolism

    7 normal subjects, 37 hour fast, four conditions including GH replacement, GH plus antilipolysis with acipimox, no GH plus antilipolysis, and GH plus antilipolysis plus intralipid. Blocking lipolysis raised urea excretion, serum urea and muscle protein breakdown by almost 50%. Adding GH during fasting with antilipolysis did not influence indices of protein degradation. Restoring free fatty acids with intralipid cut whole-body protein degradation by 10 to 15%. The authors conclude that inhibition of lipolysis eliminates the ability of GH to restrict fasting protein loss.

  5. Human2001
    The protein-retaining effects of growth hormone during fasting involve inhibition of muscle-protein breakdown

    Diabetes

    8 normal subjects, four conditions: basal postabsorptive, 40 hour fast, 40 hour fast with somatostatin suppression of GH, and 40 hour fast with GH suppressed plus exogenous GH replacement, with insulin and glucagon replaced identically in the last two arms. Urinary urea excretion in mmol per 24 hours: 392 plus or minus 44 basal, 440 plus or minus 32 fasted, 609 plus or minus 76 fasted without GH, 408 plus or minus 36 with GH replaced (P < 0.05). Forearm muscle protein breakdown moved the same way.

  6. Human1971
    Metabolic response to human growth hormone during prolonged starvation

    Journal of Clinical Investigation

    Human growth hormone administered to obese subjects during 5 to 6 weeks of starvation. Serum insulin rose 2 to 3 fold, preceding a 50% rise in blood glucose. Lipolysis persisted. Blood beta-hydroxybutyrate and acetoacetate rose 20 to 40%, giving total blood ketoacid concentrations of 10 to 12 mmol/L and ketonuria of 150 to 320 mmol per day, with increased urinary potassium loss and symptoms of nausea, vomiting, weakness and myalgias. Despite a 50% reduction in urea excretion, total nitrogen loss was unchanged because urinary ammonia rose 50%.

  7. Human2013
    Randomized cross-over trial of short-term water-only fasting: metabolic and cardiovascular consequences

    Nutrition, Metabolism and Cardiovascular Diseases

    30 apparently healthy volunteers, randomised crossover of one day of water-only fasting against a day of usual eating. Growth hormone rose (p = 1.1 x 10 to the minus 4). Haemoglobin, red cell count, haematocrit, total cholesterol and HDL rose; triglycerides, bicarbonate and weight fell. In those randomised to fast first, most markers including growth hormone and cholesterol had returned to baseline by 48 hours. The magnitude of the growth hormone change is not stated in the abstract.