The tissue that makes up your structure
The musculoskeletal system is not one tissue. It is a hierarchy of connected tissues with different compositions, different turnover rates, different vascular supplies, and different biochemical requirements. Understanding which tissue is the problem — and what that tissue needs to function, repair, and tolerate load — is the starting point for a biochemically-informed approach to musculoskeletal health.
Primary substrate: Creatine phosphate (explosive), glycogen (sustained), fatty acids (endurance)
Repair rate: Fast — satellite cells activated within hours of damage
Key nutrients: Creatine, protein (leucine threshold), magnesium, B vitamins, carnosine
Failure mode: Sarcopenia (ageing), rhabdomyolysis (overload), contracture (chronic shortening)
Primary structure: Type I collagen (tendons), mixed Type I/III collagen (ligaments)
Repair rate: Slow — poor vascular supply, weeks to months for meaningful remodelling
Key nutrients: Glycine, proline, vitamin C, copper, silicon, zinc
Failure mode: Tendinopathy — failed healing response, not inflammation as previously thought
Primary structure: Type II collagen + aggrecan proteoglycan + water (70% by weight)
Repair rate: Extremely slow — avascular, relies on synovial fluid diffusion
Key nutrients: Glycine, proline, vitamin C, glucosamine, chondroitin, omega-3
Failure mode: Osteoarthritis — progressive matrix degradation exceeding synthesis
Primary structure: Type I and III collagen with contractile myofibroblasts — fascia actively contracts
Repair rate: Moderate — better vascularised than tendons, slower than muscle
Key nutrients: Collagen precursors, magnesium (myofibroblast relaxation), omega-3
Failure mode: Fibrosis, adhesions, plantar fasciitis, frozen shoulder
Primary structure: Synovial membrane producing hyaluronic acid — the joint's lubricating fluid
Repair rate: Moderate — membrane is vascularised but HA production declines with age and inflammation
Key nutrients: Omega-3 (reduces inflammatory cytokines in synovial fluid), vitamin D, glucosamine
Failure mode: Inflammatory arthritis — immune-mediated synovial destruction
Primary structure: Myelin sheath (phospholipid) surrounding axons — myelin requires B12, choline, omega-3
Repair rate: Slow — 1–4mm per day of axon regrowth after injury
Key nutrients: B12 (methylcobalamin), B6, alpha-lipoic acid, omega-3, magnesium
Failure mode: Peripheral neuropathy, nerve entrapment, referred pain patterns
Creatine — beyond sport
Creatine monohydrate is one of the most researched sports supplements, and one of the few with consistent trial evidence of benefit (Kreider 2017). The popular picture, that creatine makes you stronger and helps you lift more, is true. The wider claims made for it, for the brain, bones and blood sugar, are weaker, and below each one is graded on what the trials found.
Creatine — what the evidence supports
One molecule, graded claim by claim
Creatine is made in the body from arginine and glycine, with a methyl group from methionine, mainly in the kidneys and liver. About 95% is stored in skeletal muscle; the remaining 5% or so is mostly in the brain and testes (Kreider 2017). As phosphocreatine it refills ATP faster than any other energy system, which is why it matters most in short, maximal efforts.
Muscle
Strength and lean mass
Resistance training is the main way to keep muscle with age. Of the supplements tested alongside it, creatine has the most consistent evidence: across 22 trials in older adults, about 1.4 kg more lean mass and slightly more strength than training alone (Chilibeck 2017).Strong, for a small extra effect on top of training
Brain
Memory and sleep loss
Meta-analyses find a small benefit to memory, with no clear effect on executive function, and they disagree on who benefits most (Prokopidis 2023; Xu 2024). Small trials suggest it blunts the drop in performance during sleep deprivation (McMorris 2006).Mixed: small effects, small trials
Brain disease
Neuroprotection: not shown
Large trials in Parkinson’s disease (1,741 people) and Huntington’s disease (553 people) were stopped early because creatine wasn’t working (Kieburtz 2015; Hersch 2017). The protective effects seen in animals haven’t carried over to people.Tested and failed in people
Bone
Bone density: not shown
One small 12-month trial in postmenopausal women slowed bone loss at the hip (Chilibeck 2015). The larger two-year trial, in 237 women, found no effect on bone density, though some measures of hip bone shape held up better (Chilibeck 2023). A 2026 meta-analysis found bone density unchanged overall (Naddafha 2026).Mixed, leaning no effect
Blood sugar
Glucose control: unproven
In one 12-week trial of 25 people with type 2 diabetes who were also exercising, creatine lowered HbA1c (Gualano 2011). Pooled trials show no effect on fasting glucose or insulin resistance (Delpino 2021).Mixed: one small positive trial
Energy
The phosphocreatine buffer
Phosphocreatine buffers ATP in high-energy tissues, including heart muscle and neurons. That is the reason for the brain and muscle research above; it isn’t, by itself, evidence of a benefit.Mechanism
Form, loading and safety
Form. Monohydrate is the form with the evidence. Ethyl ester, buffered and other forms haven’t been shown to work better, and ethyl ester did worse in a head-to-head trial (Kreider 2017; Spillane 2009; Jagim 2012).
Loading. A loading week fills the muscles faster; a lower daily amount gets to the same level in about four weeks (Hultman 1996). So loading gets you there sooner, not further.
Cycling. No trial has compared cycling with taking it continuously. Your own production isn’t suppressed after stopping, and long-term use hasn’t shown harm (Kreider 2017), so there’s no evidence-based reason to cycle.
Kidneys and blood tests. In healthy adults, trials lasting up to about a year show no change in kidney function (eGFR) or urea. Blood creatinine does rise, by about 0.13 mg/dL (about 11 µmol/L), because creatine breaks down to creatinine (Tsiaras 2026). Tell whoever orders your blood tests that you take it, so a kidney problem isn’t wrongly suspected. If you have kidney disease, ask your doctor first.
How much? What the trials used
This is what was tested, not a prescription. If you take medication, have a medical condition, are pregnant or are under 18, talk to whoever looks after you first.
Creatine monohydrate: either 5 g four times a day (about 0.3 g per kg) for 5–7 days, then 3–5 g a day; or 3 g a day, which fills the muscles more gradually, over about 28 days (Kreider 2017). Taking it with a meal containing carbohydrate, or carbohydrate and protein, helps the muscles retain more (Kreider 2017). Expect a small weight gain, mostly water in the muscle.
Who it’s worth considering for (my view, not a trial finding): anyone losing muscle, anyone over 45, and anyone in hard training. The age of 45 isn’t a threshold from the trials; it’s where I start raising it with clients.
Using creatine for sport? The four other supplements with good evidence for performance are in Sports Supplements: The Short List That Works.
Sources for the creatine section and the nutrient table
- Kreider RB et al. ISSN position stand: safety and efficacy of creatine supplementation. J Int Soc Sports Nutr 2017. PMID 28615996. doi:10.1186/s12970-017-0173-z
- Chilibeck PD et al. Creatine during resistance training in older adults: a meta-analysis. Open Access J Sports Med 2017. PMID 29138605. doi:10.2147/OAJSM.S123529
- Prokopidis K et al. Creatine and memory in healthy individuals: meta-analysis. Nutr Rev 2023. PMID 35984306. doi:10.1093/nutrit/nuac064
- Xu C et al. Creatine and cognitive function in adults: meta-analysis. Front Nutr 2024. PMID 39070254. doi:10.3389/fnut.2024.1424972
- McMorris T et al. Creatine and sleep deprivation. Psychopharmacology 2006. PMID 16416332. doi:10.1007/s00213-005-0269-z
- Kieburtz K et al (NINDS NET-PD). Creatine in Parkinson disease: randomised trial. JAMA 2015. PMID 25668262. doi:10.1001/jama.2015.120
- Hersch SM et al. CREST-E: creatine for Huntington disease. Neurology 2017. PMID 28701493. doi:10.1212/WNL.0000000000004209
- Chilibeck PD et al. Creatine and resistance training on bone in postmenopausal women (12 months). Med Sci Sports Exerc 2015. PMID 25386713. doi:10.1249/MSS.0000000000000571
- Chilibeck PD et al. A 2-year randomised trial of creatine for postmenopausal bone health. Med Sci Sports Exerc 2023. PMID 37144634. doi:10.1249/MSS.0000000000003202
- Naddafha S et al. Creatine for lean mass, strength and bone density in postmenopausal women: meta-analysis. J Int Soc Sports Nutr 2026. PMID 42141930. doi:10.1080/15502783.2026.2668435
- Gualano B et al. Creatine in type 2 diabetes: randomised trial. Med Sci Sports Exerc 2011. PMID 20881878. doi:10.1249/MSS.0b013e3181fcee7d
- Delpino FM, Figueiredo LM. Creatine, glycaemic control and insulin resistance: meta-analysis. Clin Nutr ESPEN 2021. PMID 35063192. doi:10.1016/j.clnesp.2021.11.006
- Spillane M et al. Creatine ethyl ester vs monohydrate. J Int Soc Sports Nutr 2009. PMID 19228401. doi:10.1186/1550-2783-6-6
- Jagim AR et al. Buffered creatine (Kre-Alkalyn) vs monohydrate. J Int Soc Sports Nutr 2012. PMID 22971354. doi:10.1186/1550-2783-9-43
- Hultman E et al. Muscle creatine loading in men. J Appl Physiol 1996. PMID 8828669. doi:10.1152/jappl.1996.81.1.232
- Tsiaras A et al. Creatine and kidney function: meta-analysis of randomised trials. J Ren Nutr 2026. PMID 42035842. doi:10.1053/j.jrn.2026.04.010
- Meléndez-Hevia E et al. The capacity for glycine synthesis does not satisfy the need for collagen synthesis (modelling study). J Biosci 2009. PMID 20093739. doi:10.1007/s12038-009-0100-9
- Senftleber NK et al. Marine oil supplements for arthritis pain: meta-analysis. Nutrients 2017. PMID 28067815. doi:10.3390/nu9010042
- Sigaux J et al. Type and dose of PUFA supplementation in inflammatory rheumatic diseases: meta-analysis. Arthritis Res Ther 2022. PMID 35526074. doi:10.1186/s13075-022-02781-2
- Garrison SR et al. Magnesium for skeletal muscle cramps. Cochrane Database Syst Rev 2020. PMID 32956536. doi:10.1002/14651858.CD009402.pub3
- Costello RB et al. The case for an evidence-based reference interval for serum magnesium. Adv Nutr 2016. PMID 28140318. doi:10.3945/an.116.012765
- García-Coronado JM et al. Collagen supplementation and osteoarthritis symptoms: meta-analysis. Int Orthop 2019. PMID 30368550. doi:10.1007/s00264-018-4211-5
- Shaw G et al. Vitamin C-enriched gelatin before intermittent activity augments collagen synthesis. Am J Clin Nutr 2017. PMID 27852613. doi:10.3945/ajcn.116.138594
Collagen synthesis — the structural repair pathway
Collagen is the most abundant protein in the body — approximately 30% of total protein mass. It is the primary structural component of tendons, ligaments, cartilage, bone matrix, fascia, skin, and the walls of blood vessels. The ability to synthesise and remodel collagen determines whether structural tissue repair keeps pace with damage — and whether ageing produces progressive structural deterioration or not.
Collagen synthesis is not simply a matter of eating protein. It is a multi-step enzymatic pathway with specific cofactor requirements that are frequently insufficient in people with musculoskeletal problems.
Collagen Synthesis Pathway
From amino acids to structural matrix — and where the process fails
→
Step 1: Substrate availability. Collagen is approximately 33% glycine, 11% proline, and 10% hydroxyproline by amino acid content. Dietary glycine from bone broth, connective tissue, and skin provides the most concentrated source. Glycine is conditionally essential — endogenous synthesis from serine is insufficient to meet collagen repair demands, particularly under chronic injury or in older adults. Glycine · Proline · Hydroxyproline
→
Step 2: Hydroxylation. Prolyl hydroxylase and lysyl hydroxylase convert proline and lysine to hydroxyproline and hydroxylysine — the cross-linking amino acids that give collagen its tensile strength. Both enzymes require vitamin C as an essential cofactor. Vitamin C deficiency does not produce scurvy immediately — it first impairs collagen quality long before clinical signs appear. Vitamin C (ascorbic acid)
→
Step 3: Triple helix formation. Three pro-collagen chains wind into the characteristic triple helix — the source of collagen's extraordinary tensile strength. This process requires appropriate temperature, pH, and the presence of specific chaperone proteins. Magnesium is required for procollagen processing enzyme activity. Magnesium
→
Step 4: Cross-linking. Lysyl oxidase — a copper-dependent enzyme — catalyses the cross-linking of collagen fibres into the mature matrix that provides structural integrity. Copper deficiency impairs this step. Low copper → inadequate cross-linking → structurally weak collagen regardless of how much collagen precursor is available. Copper (via lysyl oxidase)
→
Step 5: Matrix remodelling. Matrix metalloproteinases (MMPs) degrade old or damaged collagen while fibroblasts synthesise new matrix. This remodelling process responds to mechanical load — tissue stressed appropriately within its tolerance strengthens; tissue stressed excessively degrades. Zinc is required for MMP activity. Omega-3 modulates the inflammatory signals that regulate fibroblast activity. Zinc · Omega-3 · Mechanical load
Tissue tolerance — the concept that changes how you think about injury
Tissue tolerance — the key clinical concept
Injury is not an event. It is the cumulative consequence of load exceeding tolerance.
Every tissue has a tolerance threshold — the amount of mechanical, chemical, and thermal stress it can absorb before structural or functional failure occurs. Below the threshold, appropriate loading drives adaptation: tendons thicken, muscle hypertrophies, bone density increases, cartilage matrix is maintained. Above the threshold, damage accumulates faster than repair — and if this continues long enough, the tissue eventually fails. The clinical insight from tissue tolerance is that injury is almost never caused by a single event. The person who "throws their back out" lifting a bag of shopping did not injure themselves lifting the bag. The bag was the last load in a long series of sub-threshold insults that progressively eroded the tissue's reserve capacity until the margin was gone. The bag was the straw. The back had been loading toward failure for months or years. The same logic applies to tendinopathy, osteoarthritis, stress fractures, and rotator cuff tears. The acute event gets the diagnosis. The chronic loading history is the actual cause. Addressing the acute event without addressing the loading history and the biochemical environment that allowed it to develop is addressing the symptom, not the problem.
The Janda model — why muscles fail in predictable patterns
Vladimir Janda's Muscle Classification
Tonic muscles shorten. Phasic muscles lengthen. The pattern is predictable — and biochemically reinforced.
Tonic muscles — tend to shorten and become facilitated under stress
Hip flexors (iliopsoas)
Upper trapezius
Pectorals (major and minor)
Hamstrings
Piriformis
Suboccipital muscles
Quadratus lumborum
Levator scapulae
Phasic muscles — tend to lengthen and become inhibited under stress
Gluteus maximus
Deep neck flexors
Rhomboids and middle trapezius
Vastus medialis
Serratus anterior
Lower abdominals
Tibialis anterior
Peroneals
Janda's model matters clinically because it predicts where musculoskeletal dysfunction will develop before it becomes painful. Shortened hip flexors from prolonged sitting inhibit their antagonists — the glutei. Inhibited glutei fail to control hip extension, placing excessive demand on the lumbar erectors and hamstrings. Overloaded hamstrings develop tendinopathy or strain. The site of pain — the hamstring — is not the site of the problem — the hip flexor/glute relationship. Treating the hamstring without addressing the hip flexor shortening is treating downstream of the problem.
The biochemical dimension adds a layer Janda's anatomical model cannot address. Magnesium deficiency maintains tonic muscle hypertonicity — magnesium is required for muscle relaxation (calcium induces contraction, magnesium induces relaxation). A chronically tight upper trapezius in someone with low magnesium will not respond adequately to manual therapy or stretching because the biochemical driver of the tension is unaddressed. Vitamin D deficiency impairs muscle function at the fibre level — low vitamin D is associated with muscle weakness, pain, and reduced force production independent of any structural or movement problem. Omega-3 deficiency maintains synovial inflammation that perpetuates joint stiffness and pain regardless of movement quality.
The ascending pain sensitisation pathway — why chronic pain is a nervous system story
Pain Sensitisation Cascade
From tissue nociception to central sensitisation — why the pain outlasts the injury
1
Peripheral nociception
Tissue damage activates nociceptors — pain-sensing nerve endings. Inflammatory mediators (prostaglandins, bradykinin, substance P) lower the nociceptor threshold, producing hyperalgesia (increased pain sensitivity) at the injury site. This is the appropriate acute response — it protects the tissue from further loading.
2
Spinal sensitisation
Repeated peripheral nociceptive input sensitises dorsal horn neurons in the spinal cord — a process called wind-up. The synaptic threshold lowers. Stimuli that previously produced no pain now produce pain. The pain begins to spread beyond the original injury site as adjacent spinal segments become sensitised.
3
Central sensitisation
With prolonged input, central pain processing changes at the brain level. The anterior cingulate cortex and insula — areas involved in pain's emotional and cognitive dimensions — become hyperactivated. Pain becomes less about tissue state and more about nervous system state. Non-painful stimuli (light touch, temperature) produce pain. This is allodynia — and it explains why chronic pain persists long after tissue healing has occurred.
4
The 10,000-hour movement engram
Motor patterns are learned through repetition — thousands of movement repetitions groove neural pathways that become automatic. A movement pattern performed with faulty loading mechanics over years produces two problems simultaneously: cumulative tissue damage from the loading, and a deeply entrenched motor engram that perpetuates the faulty pattern even after awareness and intent to change. This is why motor re-education requires more than instruction — it requires thousands of correctly-loaded repetitions to overwrite the established pattern. Pain inhibits motor activation and further entrenches compensatory patterns. The cycle is: faulty loading → tissue damage → pain → motor inhibition → more faulty loading.
5
The biochemical amplifier
Systemic inflammation amplifies pain sensitisation at every level of the cascade. Inflammatory cytokines (IL-6, TNF-alpha) lower nociceptor thresholds, increase spinal sensitisation, and impair descending pain inhibition from the periaqueductal grey. A client with chronic musculoskeletal pain and elevated hsCRP is not just inflamed — their inflammatory state is actively maintaining and amplifying their pain experience. Addressing the systemic inflammation is addressing the pain, not just treating the tissue.
The nutritional foundations of musculoskeletal health
| Nutrient |
Evidence |
Mechanism |
Clinical priority |
| Creatine monohydrate |
Strong |
Phosphocreatine replenishment for maximal efforts; a small extra gain in lean mass and strength when added to resistance training (Chilibeck 2017) |
Trial doses in the How much box above. My view: worth considering if you’re losing muscle, over 45, or in hard training |
| Glycine |
Mechanism |
About one in three amino acids in collagen. One modelling study suggests the body may not make enough for all its needs (Meléndez-Hevia 2009); this hasn’t been tested in people |
Food sources: bone broth, skin, collagen. No trial has tested glycine supplements for collagen repair |
| Vitamin C |
Biochemistry |
Cofactor for prolyl and lysyl hydroxylase, the enzymes that make collagen strong. Severe deficiency (scurvy) breaks down connective tissue |
Food first: kiwi, peppers, citrus. No trial has set a vitamin C dose for joints or tendons. Worth checking if your diet is low in fruit and veg |
| Omega-3 (EPA+DHA) |
Moderate (rheumatoid arthritis) |
Lowers inflammatory prostaglandins and leukotrienes |
In rheumatoid arthritis, fish oil gives a small reduction in joint pain (Senftleber 2017). Not shown for osteoarthritis, and higher doses haven’t worked better (Sigaux 2022). An omega-3 index test shows your level |
| Magnesium |
Biochemistry; cramps: not supported |
Needed for muscle relaxation and for hundreds of enzymes, including ATP production |
For night cramps in older adults, magnesium probably doesn’t help (Cochrane review, Garrison 2020). Labs call 0.75–0.95 mmol/L (1.8–2.3 mg/dL) normal; some researchers argue the lower limit is too low, because it came from a 1970s population survey, not from outcomes (Costello 2016) |
| Vitamin D |
Strong for correcting deficiency |
Vitamin D receptors in muscle; severe deficiency causes muscle weakness and bone pain (osteomalacia) |
Test, and correct a deficiency. Guidelines don’t agree on an optimal level above that; see my vitamin D page for the evidence |
| Copper |
Biochemistry |
Needed for lysyl oxidase, the enzyme that cross-links collagen fibres |
Deficiency is uncommon; one known cause is long-term high-dose zinc, because zinc and copper compete for absorption. Liver is a rich food source |
| Collagen hydrolysate |
Mixed |
Supplies glycine, proline and hydroxyproline |
In osteoarthritis trials, collagen improved overall symptom scores but not the pain score on its own (García-Coronado 2019). In a small lab study, gelatin with vitamin C taken an hour before exercise raised a blood marker of collagen production (Shaw 2017); whether that means stronger tendons hasn’t been shown |
The person whose back goes into spasm every few months, who takes two weeks off, recovers, and then repeats the cycle six months later is not unlucky. They are operating at or near their tissue tolerance limit, in a biochemical environment that is insufficient for adequate repair, with a motor pattern that perpetuates the loading that caused the problem. Addressing any one of these three without the others is addressing a third of the problem.
What functional testing adds to the musculoskeletal picture
Most musculoskeletal presentations are assessed through imaging (X-ray, MRI, ultrasound) and clinical examination. What these do not assess is the biochemical environment in which the tissue is trying to function and repair. Vitamin D deficiency producing genuine myopathy can mimic fibromyalgia, non-specific low back pain, and generalised joint pain — and resolves completely with normalisation of vitamin D. Magnesium insufficiency producing chronic muscle hypertonicity is treated with manual therapy and stretching that provides temporary relief until the magnesium level is addressed. Systemic inflammation elevating the pain sensitisation threshold means manual therapy, exercise, and analgesics are working against a headwind that the anti-inflammatory picture could resolve.
The Regenerus Functional Platinum blood panel covers vitamin D, magnesium, hsCRP, ferritin (iron for myoglobin), and CK (creatine kinase — an indicator of muscle stress and damage). The Metabolomix+ OAT covers vitamin C (ascorbic acid), B vitamins relevant to energy metabolism in muscle, and oxidative stress markers. Omega-3 index on the fatty acids add-on quantifies EPA and DHA directly. Together these give a comprehensive biochemical picture of the musculoskeletal environment that no imaging study provides.
Know the biochemistry behind your structural health
Vitamin D, magnesium, omega-3 index, vitamin C, inflammatory markers, and creatine kinase — the functional blood picture that tells you whether your tissue has the substrates to repair and the biochemical environment to heal.
Book a conversation →
Related reading