Tendon Loading: How the Achilles Adapts, and How It Gets Hurt

A literature summary for the gymnastics injury-prevention Achilles tendon model

The headline number

A tumbling pass does ~520× the fatigue damage of one heavy slow resistance rep, while delivering ~2% of the adaptive signal.

That is the central output of the consortium’s Achilles tendon model, and it follows from a well-measured piece of physiology: tendon is viscoelastic, and a ~130 ms landing is over before the tissue can transmit the deformation to the cells that would adapt to it. Bohm and colleagues held tendon strain constant and varied only repetition duration — 3-second reps produced +57% stiffness over 14 weeks; 130 ms jump landings produced essentially nothing (Bohm et al. 2014, J R Soc Interface — a separate study from the meta-analysis cited as [1] below). So a landing-dominated training week loads the tendon hard enough to accumulate fatigue damage, but not in a way that tells it to get stronger.

One honest caveat: the ratio is a model estimate, not a measurement — no study has ever measured Achilles tendon force during gymnastics, so the tumbling inputs are extrapolated from ground reaction forces. Treat the ~520× as directional. The direction, however, is anchored on both sides by the human experiments summarized below.

1. Purpose and scope

This document summarizes the highest-quality available evidence on two related questions: how mechanical loading makes tendon stronger (adaptation), and how mechanical loading damages tendon (fatigue injury and tendinopathy). It draws on 16 papers chosen as the strongest available sources on these topics — systematic reviews and meta-analyses where they exist, landmark conceptual reviews, and seminal primary experiments where a review doesn’t cover the specific quantitative parameter needed. The goal is to give the Achilles tendon computational model a set of literature-backed numbers — strain thresholds, time courses, cycle counts — to check its behavior against; Section 6 pulls these together for that purpose, and Section 7 translates the findings into gym-floor language for coaches without a science background. Each science section ends with a short “in plain English” recap. Ten of the 16 papers were downloaded as PDFs, stored in the publications/ folder; six are paywalled and summarized here from their abstracts and citing literature, flagged in Section 8.

2. How tendon responds to load: the basics

A tendon is a rope of collagen fibers that transmits force from muscle to bone. Two properties describe how it behaves mechanically: stiffness (how much the tendon stretches for a given force — a whole-tendon, geometry-dependent property) and its close cousin, the elastic modulus (essentially stiffness normalized for the tendon’s size, so it reflects the quality of the material itself rather than how much of it there is). A third property, cross-sectional area (CSA), is simply how thick the tendon is — its “cross-section,” like a slice of rope.

Tendon is not inert rope, though — it is living tissue. Cells embedded in it, called tenocytes, sense mechanical strain (the percentage the tendon stretches under load) and convert that mechanical signal into a biochemical one, a process called mechanotransduction [8]. That signal controls two competing processes: collagen synthesis (building new structural protein) and collagen degradation, driven largely by enzymes called matrix metalloproteinases (MMPs) that break collagen down. Loading turns up both processes at once — synthesis and degradation both increase with mechanical loading — and whether the tendon nets out stronger or weaker depends on the balance between them over time [6].

One finding reshapes how that balance should be understood mechanistically. Using a clever technique — measuring residual radioactive carbon-14 from 1950s–60s nuclear bomb tests, present in different amounts in the atmosphere (and therefore in newly built tissue) year by year — Heinemeier and colleagues showed that the core collagen of an adult human Achilles tendon is essentially never replaced after growth stops, around age 17 [4]. The tendon core you have as an adult is largely the same collagen you built as a teenager, strikingly unlike skeletal muscle (sampled in the same study), which turns over collagen continuously throughout adult life. The practical implication: adult tendon adaptation to training cannot mean “the collagen gets replaced with stronger collagen.” It more likely means new collagen and cross-links added to the existing structure, changes in how existing fibers are organized and packed, and new material added at the periphery — not wholesale renewal of the load-bearing core [4]. For a model, this argues against treating adult tendon as a simple turnover/replacement compartment, and suggests that adolescent tendon (still actively laying down its permanent core) may respond to loading differently than adult tendon does [3,4].

In plain English: A tendon is a rope of collagen connecting muscle to bone, with living cells inside that feel every stretch. Those cells are constantly both building new rope material and breaking old material down — training shifts the balance. The big surprise: the core of an adult’s Achilles is built during the teenage years and then never replaced for the rest of life. Adults can renovate their tendon, but only teenagers are still building it — which means a teenage gymnast is literally constructing the tendon they will compete on forever.

3. Loading that drives adaptation

Across the four adaptation meta-analyses/reviews in this set, one finding stands out consistently: strain magnitude — how much the tendon stretches under load, as a percentage of its resting length — is the single most important loading parameter for driving adaptation. It matters more than training volume, more than intensity relative to muscle strength, and, surprisingly, more than the type of muscle contraction used to load the tendon.

The most direct evidence for a strain threshold comes from a controlled human trial by Arampatzis and colleagues [7]. In a within-subject design, one leg trained at low tendon strain (2.85% ± 0.99%) and the other at high tendon strain (4.55% ± 1.38%), matched for how often and how much loading was applied, over 14 weeks. Only the high-strain leg showed increased tendon stiffness, increased elastic modulus, and measurable thickening (hypertrophy); the low-strain leg showed no adaptation beyond ordinary daily activity. The authors interpret this as evidence of a strain “set point”: the tendon appears to target a habitual strain level during normal activity, and loading must exceed that set point by a meaningful margin — somewhere between 2.85% and 4.55% strain here — before it triggers adaptive remodeling [7]. This is the single most specific strain-threshold number in the literature reviewed here.

The two large meta-analyses of human training studies support the same conclusion at a population level. Bohm and colleagues, pooling 37 exercise interventions across 264 people, found tendon stiffness and modulus reliably increase with training (stiffness effect size SMD=0.70, modulus SMD=0.69 — SMD, or standardized mean difference, is a common way to size an effect across studies using different measurement scales; roughly 0.2 is small, 0.5 medium, 0.8 large), and that adaptation depended significantly on loading intensity/magnitude but not on whether training used isometric, concentric, or eccentric contractions — those produced comparable results once matched for strain [1]. The larger 2022 update by Lazarczuk and colleagues (61 studies, 763 participants) sharpened this: higher-strain protocols produced significantly greater gains in both modulus (SMD=0.82 vs. lower-strain protocols, p=0.009) and stiffness (SMD=1.04, p=0.007) [3]. Their overall pooled effects were stiffness SMD=0.74, modulus SMD=0.82 (large), and CSA SMD=0.22 (small) [3] — consistent with Bohm’s numbers and confirming that loading changes the quality of tendon material more than its size, at least over the weeks-to-about-a-year timescale these studies cover.

That last point — material change outpacing size change — is itself model-relevant. Wiesinger and colleagues’ systematic review found that short- and medium-term training (most studies 8–14 weeks) reliably changes stiffness and modulus, but shows no clean, steadily-increasing dose-response within that timeframe — individual responses vary widely [2]. Longer-term training (years, not weeks) is instead associated with larger CSA (actual thickening), without much further material change. Wiesinger proposes a two-stage model: material properties adapt first, over weeks to months, while CSA is the “ultimate adjusting parameter” that dominates only after years of training [2] — suggesting a computational model may need two adaptation time constants, a fast one for stiffness/modulus and a much slower one for CSA, rather than a single curve.

On duration: interventions of at least 8 weeks reliably produced detectable adaptation in the Bohm meta-analysis, with a non-significant trend toward larger effects at 12+ weeks [1]. Loading intensity above roughly 70% of maximum voluntary contraction was associated with larger adaptation in the same analysis [1]. One caveat for a youth-athlete model: Lazarczuk’s pooled data found larger adaptation magnitude in adults than in children/adolescents [3] — plausibly linked to adolescent tendon still actively building its permanent core structure [4], so its response to loading may not follow adult rules.

In plain English: Tendons only get stronger when they’re loaded hard — hard enough to stretch them well beyond what everyday activity does. How heavy the load is matters far more than which exercise you pick or how many reps you do; light, easy work does essentially nothing to the tendon, no matter how much of it you do. Change takes a minimum of about two months of consistent heavy loading. And the tendon improves its quality (springiness, stiffness) within weeks-to-months, but only grows visibly thicker after years of training.

4. Loading that causes damage

Damage evidence tells a story that runs in parallel to the adaptation evidence, sharing many of the same mechanisms but tipping toward net breakdown instead of net gain.

The most quantitative dosing data comes from an in vivo rat model that cyclically loaded patellar tendons at about 40% of their failure strength, producing an initial peak cyclic strain around 4.5% ± 0.6% — coincidentally close to the Arampatzis “high-strain, adaptive” threshold above, underlining that similar strains can drive either adaptation or damage depending on the loading pattern [13]. Fung and colleagues defined three graded levels of fatigue damage (Low, Moderate, High) by how much peak strain increased beyond a stable early plateau: reaching Low-level damage took roughly 10,600 cycles on average, Moderate/High roughly 17,000 cycles, with substantial variability between animals [13]. Mechanically, stiffness dropped about 18% at Low and Moderate damage; at High damage, hysteresis (energy lost per loading cycle, a measure of internal friction) instead increased about 25%, suggesting a shift in the failure mechanism itself rather than just more of the same damage [13]. Structurally, damage progressed from localized “kinked” fibers (mild), to kinking plus fiber delamination (moderate), to fiber angulation and outright discontinuity (severe) [13] — a staging that lines up remarkably well with the clinical “continuum” model described below.

That clinical continuum model, from Cook and Purdam, organizes tendon pathology into three stages driven by cumulative load: reactive tendinopathy (a short-term, non-inflammatory thickening response to a sudden overload — potentially reversible if load is reduced), tendon disrepair (an attempted-healing state with more matrix breakdown and disorganization, from chronic overload), and degenerative tendinopathy (further progression with cell death and matrix disorganization that does not reverse) [9]. A tendon can move backward toward normal in the early, reactive stage if load is modified, but degenerative regions do not revert to normal tissue [9]. A 2016 update, and a related imaging study, add two qualifiers: tendon pain correlates poorly with structural damage visible on imaging (a degenerated-looking tendon can be pain-free, and a near-normal one can hurt) [10]; and even symptomatic tendons typically retain a large amount of normally organized tissue around a focal area of damage — the basis of the “treat the donut, not the hole” principle, which argues for building load tolerance in the surrounding healthy tissue rather than trying to normalize the damaged core [16].

Direct causal evidence that overuse — not just correlation — produces mechanical weakening comes from a classic rat overuse (downhill treadmill running) study of the supraspinatus tendon (a shoulder tendon, used here as a general overuse model since no Achilles-specific data of this design exist). Sustained overuse over 4, 8, and 16 weeks progressively reduced elastic modulus (down to 52–61% of control) and stress at failure (down to 51–63% of control), while CSA paradoxically increased (129% of control at 4 weeks, 164% by 16 weeks) — a disorganized, weaker, but thicker tendon, consistent with the reactive/disrepair stages above [12].

Energy-storing tendons — tendons like the Achilles and patellar tendon that stretch and recoil elastically to store and return energy during locomotion, unlike “positional” tendons that mostly just transmit force — are a specifically higher-risk category. Using an equine analog (the superficial digital flexor tendon, functionally like a human Achilles), Thorpe and colleagues measured in vivo strains during normal galloping in excess of 10%, up to ~16% — far above the ~4.5% strain used to fatigue-damage tendon in the rat model above, showing energy-storage tendons routinely operate close to injury-relevant strain levels even during normal use [14]. Consistent with that risk, human Achilles tendinopathy affects roughly 3% of the general population but 15–56% of elite athletes, depending on sport [14] — risk clearly elevated by high-strain, energy-storage athletic loading, directly relevant to gymnastics.

Finally, damage is not only a story of overloading. Arnoczky and colleagues showed that when local tendon fibers lose their normal mechanical connection to nearby cells — for instance, because a micro-injury “shields” adjacent tenocytes from strain — those under-stimulated cells respond by increasing production of collagen-degrading enzymes, not decreasing it [15]. Cells that sense too little strain locally can drive matrix breakdown just as cells that sense too much strain can. This “stress-shielding” pathway means a purely damage-above-a-strain-threshold model is likely incomplete: a small area of existing damage may propagate on its own even without further overload, and complete rest/immobilization is not a risk-free response to injury [15,6].

In plain English: Tendon damage works like bending a paperclip — no single bend breaks it, but thousands of hard bends slowly do. Trouble comes in stages: first the tendon gets “cranky” and thickens (this stage is fixable by easing off), then, if overload continues, it moves through worsening disorganization to a permanently degenerated state that doesn’t heal back to normal. Two traps to know about: pain is a poor guide to how damaged a tendon actually is (badly degenerated tendons can feel fine, and mildly affected ones can hurt a lot), and doing nothing is also harmful — a completely rested or immobilized tendon actively weakens rather than staying the same.

5. The adaptation–damage balance

The same load, at the same strain magnitude, can adapt or damage a tendon — the outcome depends on dose, progression, and rest, not strain magnitude alone. Mechanical loading acutely turns up both collagen synthesis and collagen-degrading enzyme activity together [6]; whether a training bout leaves the tendon net stronger or net weaker depends on the relative timing of these two responses. Synthesis is reported to peak roughly 24 hours after a loading bout and stay elevated for about three days, while degradation peaks earlier [11]. If further loading occurs before synthesis has caught up, the balance can tip toward net breakdown — the leading candidate mechanism for why insufficient recovery between sessions, more than any single load spike, drives cumulative damage [11]. (This timing estimate comes from a paywalled source and should be treated as order-of-magnitude, worth a second check if it becomes central to the model.)

Energy-storage loading — the tumbling, landing, and rebounding loads central to gymnastics — sits at the sharpest point of this balance. It is simultaneously the type of loading shown to produce the largest adaptive stiffness gains (strain magnitude is the dominant driver of adaptation [1,3,7]) and the type shown to carry by far the highest tendinopathy risk in athletic populations (15–56% vs. 3% baseline) [14]. This is not a contradiction — the same mechanism (high strain drives a strong cellular response) produces opposite outcomes depending on whether the tendon gets adequate recovery and progressive, rather than sudden, load increases. Underloading is not a safe default either: both cellular stress-deprivation [15] and general inactivity/immobilization [6] actively drive degradative processes rather than simply pausing them, and degenerative structural change, once established, does not reverse just because load is removed [9,10]. A 2019 review by Magnusson and Kjaer is specifically dedicated to this loading/unloading/ageing/injury balance and is worth obtaining in full if the model’s rest and detraining behavior needs deeper grounding than the sources above provide [5].

In plain English: The exact same tumbling pass can make a tendon stronger or slowly wreck it — what decides the outcome isn’t the load itself, but how fast the load was ramped up and how much recovery came between hard days. After a hard session, the tendon spends roughly the next two to three days rebuilding; stack hard, springy sessions closer together than that, over and over, and breakdown outpaces repair. The springy loads at the heart of gymnastics — tumbling, leaping, rebounding — are simultaneously the best tendon-builder and the biggest tendon risk. And “just rest it” isn’t the safe option it sounds like: unloaded tendons weaken.

6. Implications for an Achilles model in gymnastics

The literature above suggests a small set of concrete anchors a model’s behavior can be checked against. These come from a modest number of studies (often one), several from animal models rather than humans, and none from Achilles-specific gymnastics data — treat them as order-of-magnitude sanity checks, not calibrated constants.

• Adaptive strain threshold: ~2.85% strain shows no measurable adaptation over 14 weeks; ~4.55% strain produces stiffness, modulus, and mild hypertrophy gains over the same period [7]. A model’s “adaptation trigger” should sit somewhere in this band, not at zero.
• Damage-onset strain and cycle count: in a rat model, ~4.5% peak cyclic strain at ~40% of failure load produced measurable fatigue damage after roughly 10,000–17,000 cycles [13]. This overlaps the adaptive-threshold strain above — outcome at this strain level should depend on cycle count and rest, not strain alone.
• Physiological strain ceiling for energy-storage tendon: normal athletic use reaches 10–16% strain [14], well above both figures above — gymnastics-relevant loading operates in a range the fatigue-damage literature has barely characterized directly.
• Recovery/rest time constant: collagen synthesis after a loading bout peaks around 24 hours and stays elevated roughly 3 days; degradation peaks earlier [11]. This suggests a minimum beneficial inter-session recovery window on the order of 48–72 hours — the least well-verified number here (paywalled primary source).
• Two-speed adaptation: material properties (stiffness, modulus) adapt over weeks to months; CSA (size) adapts over years [2]. A single adaptation time constant will misrepresent one of these two processes.
• Underloading/detraining decay: immobilization sharply reduces collagen turnover [6], and stress-shielding near existing damage actively increases degradative enzyme activity rather than just halting adaptation [15] — “too little load” should probably actively drive a damage-like state, not just freeze adaptation at zero.
• Youth caveat: adaptation magnitude is larger in adults than adolescents in the pooled data [3], and adolescent tendon core collagen is still being laid down, not yet “locked in” [4]. Since gymnasts are frequently adolescent, this is an open question rather than a settled parameter — treat adult dose-response numbers cautiously.

These are literature anchors, not validated parameters — a sensible next step would be checking whether the model’s simulated stiffness/damage trajectories fall within these ranges under comparable loading, rather than reproducing any single number exactly.

In plain English: This section collects the actual numbers from the studies — how much stretch triggers adaptation, how many hard cycles cause wear, how long recovery takes — so the computer model’s behavior can be sanity-checked against them. They’re rough anchors from a handful of studies (some in animals), not precise settings.

7. What this means in the gym: a coach’s translation

The studies above measure things no coach can see on the floor — percent strain, collagen synthesis rates, cycles to fatigue damage. This section translates them into practice. One honest caveat first: this translation goes one step beyond what the studies directly tested. The research measured tendon responses in labs; the gym-floor advice below is the reasonable practical reading of those findings, not a set of proven prescriptions — and it’s education, not medical advice for any individual athlete.

1. Tendons only respond to heavy work — light conditioning doesn’t count. The famous finding that ~4.5% tendon stretch drives adaptation while ~2.9% does nothing [7] sounds unusable, but it translates simply: tendon stretch tracks the force the muscle puts through it, and only genuinely heavy efforts stretch the tendon enough to change it. In practice, that means strength work at loads around 70% or more of the athlete’s maximum [1] — for the Achilles, think slow, heavy calf-raise work (heavy enough that 5–8 controlled reps is a real effort, about 3 seconds up and 3 seconds down), a few sets, two to three times a week. Band work, high-rep bodyweight conditioning, and easy jogging sit below the threshold — fine for other purposes, but they will not build tendon. Contraction style barely matters [1,3], so pick whichever heavy exercise the athlete can do with good control.

2. Judge a tendon program in months, not weeks. Measurable change takes about 8–12 weeks of consistent loading [1,3], and visible thickening takes years [2]. A tendon-strength block abandoned after three weeks “because nothing happened” was never given a chance.

3. The enemy is sudden change, not hard work. Fatigue damage accumulates over thousands of hard, springy cycles [13], and tendons cope well with high loads they are accustomed to. The classic danger moments are rapid changes: the first week back after a holiday or injury layoff, a new floor or surface, a sudden jump in tumbling or dismount volume before a competition. The practical tool is counting — track roughly how many jumps, tumbling passes, and dismounts each athlete does per session, and change that number gradually rather than doubling it in a week. After any break, rebuild volume over weeks, starting well below where the athlete left off.

4. Space out the springy days. After a hard session, the tendon’s rebuilding process takes roughly 2–3 days to complete [11]. Back-to-back days of maximal tumbling, leaping, and rebounding on the same legs give breakdown a head start on repair. Where the schedule allows, follow a heavy tumbling or vault day with lower-rebound work — bars, choreography, flexibility, or the heavy slow strength work from point 1, which loads the tendon without the fatigue cost of thousands of impact cycles. (This timing number is the least certain in this document — see Section 5 — but “don’t stack maximal rebound days” is a low-cost hedge.)

5. A sore tendon needs its load changed, not removed. Complete rest actively weakens tendon [6,15], and early-stage “cranky tendon” is the reversible stage [9]. The evidence-consistent response to a tendon that’s flaring: sharply reduce the springy, high-energy loads (tumbling, leaping, sprinting) for a period, keep the heavy slow strength work, and reintroduce rebound loads gradually once things settle. Pushing straight through repeated flare-ups is how tendons migrate toward the degenerative stage that doesn’t come back [9]. Persistent or worsening pain is a job for a clinician, not a training tweak.

6. Use the next-morning test, not the scan. Pain and stiffness the morning after training is the most practical everyday signal that yesterday’s dose was too much for today’s tendon — respond by trimming the springy volume, not necessarily by stopping. Imaging is a poor guide in both directions: degenerated-looking tendons can be pain-free and functional, and near-normal ones can hurt [10]. Manage the load and the symptoms, don’t chase the picture.

7. Teenage gymnasts are building their lifetime tendon. The tendon core is laid down before roughly age 17 and never replaced afterward [4], and adolescents don’t reliably follow adult dose-response rules [3]. That argues for taking the “gradual change” rules above more seriously with youth athletes, especially through growth spurts — not for loading them like small adults.

8. References and reading guide

Numbered references below; each entry lists the PDF filename in tendon-loading-research/publications/ when downloaded, or notes it as paywalled. Start-here papers: for the adaptation side, [3] (Lazarczuk 2022) is the most current and comprehensive meta-analysis; for the damage side, [9] (Cook & Purdam 2009) is the conceptual model that ties the rest of the damage literature together; for the underlying biology, [4] (Heinemeier 2013) is short, readable, and reshapes how “adaptation” should be understood mechanistically.

Adaptation

1. Bohm S, Mersmann F, Arampatzis A. “Human tendon adaptation in response to mechanical loading: a systematic review and meta-analysis of exercise intervention studies on healthy adults.” Sports Medicine – Open. 2015;1:7. DOI: 10.1186/s40798-015-0009-9. — Bohm2015_TendonAdaptationMetaAnalysis.pdf
2. Wiesinger H-P, Kösters A, Müller E, Seynnes OR. “Effects of Increased Loading on In Vivo Tendon Properties: A Systematic Review.” Medicine & Science in Sports & Exercise. 2015;47(9):1885-1895. DOI: 10.1249/MSS.0000000000000603. — Wiesinger2015_IncreasedLoadingTendonProperties.pdf
3. (start here) Lazarczuk SL, Maniar N, Opar DA, Duhig SJ, Shield A, Barrett RS, Bourne MN. “Mechanical, Material and Morphological Adaptations of Healthy Lower Limb Tendons to Mechanical Loading: A Systematic Review and Meta-Analysis.” Sports Medicine. 2022;52(10):2405-2429. DOI: 10.1007/s40279-022-01695-y. — Lazarczuk2022_LowerLimbTendonAdaptationsMetaAnalysis.pdf
4. (start here) Heinemeier KM, Schjerling P, Heinemeier J, Magnusson SP, Kjaer M. “Lack of tissue renewal in human adult Achilles tendon is revealed by nuclear bomb 14C.” FASEB Journal. 2013;27(5):2074-2079. DOI: 10.1096/fj.12-225599. — Heinemeier2013_LackTissueRenewalAchillesTendon.pdf
5. Magnusson SP, Kjaer M. “The impact of loading, unloading, ageing and injury on the human tendon.” Journal of Physiology. 2019;597(5):1283-1298. DOI: 10.1113/JP275450. — (paywalled — not in folder; summarized here from abstract only)
6. Kjaer M. “Role of extracellular matrix in adaptation of tendon and skeletal muscle to mechanical loading.” Physiological Reviews. 2004;84(2):649-698. DOI: 10.1152/physrev.00031.2003. — (paywalled — not in folder; summarized here from abstract only)
7. Arampatzis A, Karamanidis K, Albracht K. “Adaptational responses of the human Achilles tendon by modulation of the applied cyclic strain magnitude.” Journal of Experimental Biology. 2007;210(Pt 15):2743-2753. DOI: 10.1242/jeb.003814. — (paywalled — not in folder; summarized here from abstract/secondary sources — this is the source of the 2.85%/4.55% strain figures central to Section 3 and Section 6, so worth obtaining the full text if the model leans heavily on that number)
8. Wang JH-C. “Mechanobiology of tendon.” Journal of Biomechanics. 2006;39(9):1563-1582. DOI: 10.1016/j.jbiomech.2005.05.011. — (paywalled — not in folder; summarized here from abstract only)

Damage / tendinopathy

9. (start here) Cook JL, Purdam CR. “Is tendon pathology a continuum? A pathology model to explain the clinical presentation of load-induced tendinopathy.” British Journal of Sports Medicine. 2009;43(6):409-416. DOI: 10.1136/bjsm.2008.051193. — Cook2009_TendinopathyContinuum.pdf
10. Cook JL, Rio E, Purdam CR, Docking SI. “Revisiting the continuum model of tendon pathology: what is its merit in clinical practice and research?” British Journal of Sports Medicine. 2016;50(19):1187-1191. DOI: 10.1136/bjsports-2015-095422. — Cook2016_ContinuumRevisited.pdf
11. Magnusson SP, Langberg H, Kjaer M. “The pathogenesis of tendinopathy: balancing the response to loading.” Nature Reviews Rheumatology. 2010;6(5):262-268. DOI: 10.1038/nrrheum.2010.43. — (paywalled — not in folder; summarized here from abstract only — this is the source of the ~24h/~3-day synthesis-timing figure used in Sections 5 and 6, flagged there as needing double-checking)
12. Soslowsky LJ, Thomopoulos S, Tun S, Flanagan CL, Keefer CC, Mastaw J, Carpenter JE. “Neer Award 1999: Overuse activity injures the supraspinatus tendon in an animal model: A histologic and biomechanical study.” Journal of Shoulder and Elbow Surgery. 2000;9(2):79-84. DOI: 10.1067/mse.2000.101962. — Soslowsky2000_SupraspinatusOveruseModel.pdf
13. Fung DT, Wang VM, Andarawis-Puri N, Basta-Pljakic J, Li Y, Laudier DM, Sun HB, Jepsen KJ, Schaffler MB, Flatow EL. “Early response to tendon fatigue damage accumulation in a novel in vivo model.” Journal of Biomechanics. 2010;43(2):274-279. DOI: 10.1016/j.jbiomech.2009.08.039. — Fung2010_TendonFatigueDamageAccumulation.pdf
14. Thorpe CT, Riley GP, Birch HL, Clegg PD, Screen HRC. “Fascicles from energy-storing tendons show an age-specific response to cyclic fatigue loading.” Journal of the Royal Society Interface. 2014;11(92):20131058. DOI: 10.1098/rsif.2013.1058. — Thorpe2014_FascicleFatigueEnergyStoringTendons.pdf
15. Arnoczky SP, Lavagnino M, Egerbacher M. “The mechanobiological aetiopathogenesis of tendinopathy: is it the over-stimulation or the under-stimulation of tendon cells?” International Journal of Experimental Pathology. 2007;88(4):217-226. DOI: 10.1111/j.1365-2613.2007.00548.x. — Arnoczky2007_OverUnderStimulationTenocytes.pdf
16. Docking SI, Cook J. “Pathological tendons maintain sufficient aligned fibrillar structure on ultrasound tissue characterization (UTC).” Scandinavian Journal of Medicine & Science in Sports. 2016;26(6):675-683. DOI: 10.1111/sms.12491. — (paywalled — not in folder; summarized here from abstract/secondary sources)