Tennis Elbow and Its Cousins

Musculoskeletal · §171

Tennis elbow, golfer's elbow, and the wrist pain that shows up in new parents and barista jobs all share a story: a tendon got loaded faster than it could remodel, and the fix is not rest and not a cortisone shot — it is loading the same tendon in a graded way until it catches up. The cortisone shot wins the first month and loses the year — three large randomised trials show worse one-year outcomes in the injected arms than in people who did nothing or did exercises.

Respond · Course Evidence Moderate Chapter Musculoskeletal

Grip drops a fifth to two-fifths, kettles get heavy, the backhand or the screwdriver becomes the thing you avoid — and most people get sent for the shot that the trials say leaves them worse off a year out. The protocol that actually works is unglamorous: a five-to-ten-minute daily set of specific loaded exercises with a cheap resistance bar, a small dumbbell, and a 2–3 month timeline. De Quervain's at the wrist is the one case where a single well-placed cortisone injection earns its place.

The name on the diagnosis is misleading. Tennis elbow, golfer's elbow, and tenosynovitis all suggest inflammation — the "-itis" ending is the medical word for it — and inflammation is what most over-the-counter advice tries to fight. Look at the tissue under a microscope in someone who has had the pain for more than a few weeks and inflammation is mostly absent. What you find instead is a tendon that has lost its tight parallel collagen pattern, gained sloppy ground substance, and grown small new blood vessels into a place that should not have them Khan et al. 2002. The disease is the tendon trying to repair faster than it can keep up with — a build-up problem, not an attack.

The everyday version: a tendon is a rope that connects muscle to bone. It can take a lot of pull, but only if the rope is given time to remodel between hard pulls. Pick up a screwdriver thousands of times a week, hit a thousand backhands across a tournament weekend, hold a newborn in the same thumb-out position for hours — and at some point the rope's remodelling falls behind. Pain starts where the rope meets the bone (the lateral elbow bump for tennis elbow, the inner elbow bump for golfer's elbow, the thumb-side of the wrist for De Quervain's) and gets worse with the very action that built up the debt.

The clinical model that drives the modern protocol is the continuum: tendons cycle through a reactive phase (short-term, fully reversible with relative offload), a disrepair phase, and a degenerative phase (limited reversibility but capable of going quiet) Cook & Purdam 2009. The implication that matters: you cannot anti-inflame your way out of this once it has set in. You have to give the tendon a different stimulus — graded loading — and the time to use it.

What the trials actually say about cortisone

The most important finding for upper-limb tendinopathies is the one most patients never hear. Three large randomised trials all comparing the cortisone shot, exercises, and doing nothing for tennis elbow agree on the shape of the curve: cortisone wins the first six weeks and loses the year.

The systematic review across upper- and lower-limb tendinopathies tells the same story: short-term win, worse intermediate and long-term outcomes versus other treatments or no treatment Coombes et al. 2010. Whatever cortisone is doing to the tendon — likely a combination of analgesia that lets you load the tendon back into trouble, plus direct effects on collagen synthesis — it does not make the tendon better at its job.

What works instead

Loaded exercise — and the best-studied way to do it is eccentric training, the slow lengthening of a muscle under load. The version best tested in tennis elbow is the Tyler twist, a slow eccentric wrist extension done with a flexible rubber bar — three sets of fifteen reps, once a day. Adding it to standard physiotherapy beat physiotherapy alone in a 7-week trial: large drops in pain and large gains in function Tyler et al. 2010. A systematic review of eccentric exercise for tennis elbow across twelve studies confirmed the direction even when effect sizes varied Cullinane et al. 2014. The international tendinopathy consensus in 2019 endorsed progressive loading as the cornerstone treatment across tendinopathies including the elbow Vicenzino et al. 2020.

Isometric loading — pushing or holding without movement — does something different and useful. In patellar-tendon patients, a single set of holds reduced pain immediately and for hours afterward, with a measurable change in the brain's motor cortex behind it Rio et al. 2015. The same trick is widely used at the elbow when full-range exercises are still too sore — start with holds, graduate to eccentrics.

The De Quervain's exception

The wrist condition is the one place cortisone earns its keep. The Cochrane review pooled the available trials of cortisone injection in De Quervain's and found about five in six patients cured at six weeks with injection alone, versus one in seven with splinting alone — a big effect, replicated across the small literature Peters-Veluthamaningal et al. 2009. The reason for the difference is anatomical: De Quervain's is partly a sheath-thickening problem, and the cortisone has somewhere mechanical to act. Even here, the rule is one well-placed injection — not three.

The natural history

Roughly eight or nine in ten people with tennis elbow get better within one to two years regardless of what they do Sayegh et al. 2015 Cutts et al. 2020. Treatment's job is to shorten that window, reduce how bad it gets along the way, and keep you out of the chronic 10–20% who do not recover on their own.

What "keep aggravating it" actually looks like

The first few weeks are bearable — a sharp pinch on the outside of the elbow when you lift the kettle, a deeper ache by Thursday afternoon, gone by Sunday. Then the floor moves. The kettle stays painful. The doorknob you turn a hundred times a day becomes the thing you switch hands for. Shaking someone's hand becomes the thing you wince through. Your grip strength on the affected side is now somewhere between a fifth and two-fifths down — partner notices when you can't open the jar, kid notices when you can't lift them with that arm without setting your jaw.

If it is medial — golfer's elbow — the same story plays on the inside of the elbow and the forearm flexors, and the action that starts it is the screwdriver, the kettlebell swing, the pull-up. If it is De Quervain's, the wrist on the thumb side is the part that goes; lifting your infant, your phone, your bag of groceries in the wrong direction sends a stabbing pain up the back of the thumb. The postpartum version is so well-known it has a nickname.

The trajectory if nothing changes: most months, the symptoms are tolerable but specific tasks become forbidden — the racquet sport you played twice a week, the workshop hours, the kind of work the hand was for. Six months in, you start sleeping on the other side because you wake up every time you roll onto the affected arm. By the end of year one most people are better — the natural history is forgiving for the majority Sayegh et al. 2015. The minority that stay chronic past two years are mostly the ones who kept aggravating the same load (job, sport, instrument) without modifying it. And the people who took the cortisone shot at month three and went back to full volume the next week — the trials say they are sitting in a worse spot at month twelve than the people who did nothing Coombes et al. 2010.

The thing nobody warns you about: the second flare is faster and meaner than the first. You learn the early warning sign within a couple of cycles, and then the question becomes whether you have a protocol you trust enough to start it on day one.

The loaded-exercise protocol that the trials are built on

The shape of the protocol is the same across the three conditions: identify and turn down (not switch off) the activity that started it, then load the affected tendon in a slow, controlled way once a day for two to three months. Cheap, unglamorous, and the thing the trials actually support.

The progression rule that ties it together: load to the highest level the tendon tolerates without next-morning pain above about a 4 out of 10. Pain during the session is allowed; pain that climbs over consecutive sessions is not. Step up the resistance bar one grade or the dumbbell one weight when the current load stops feeling like work Vicenzino et al. 2020.

Counterforce straps (the band that sits an inch or two below the elbow) take some load off the tendon during use and are useful as a short-term pain reducer while you do the exercises. They do not heal anything on their own.

When the loading approach is not the right call

The protocol assumes the diagnosis is correct. A few patterns argue for getting a clinician's eyes on the arm before you start.

Get checked first if any of these are true:

A sudden tearing sensation when the pain started, or you cannot actively use the arm — this points at a true tendon rupture, not tendinopathy.
Pain that wakes you at night regardless of how you lie, fever, or unexplained weight loss — these are not overuse.
Numbness, pins-and-needles, or weakness in a finger or thumb pattern — the diagnosis may be a nerve entrapment, not a tendon.
Pain that sits a few centimetres further down the forearm than the elbow bump, or that fires when you resist your middle finger straightening — radial-tunnel syndrome mimics tennis elbow and is treated differently.
Multiple joints sore, morning stiffness lasting an hour, or family history of rheumatoid or psoriatic arthritis — inflammatory arthritis presents in tendons too and needs a different workup.

Two situational notes. In pregnancy and breastfeeding, the cortisone injection that the De Quervain's literature supports is usually held off — the condition often eases when the constant lifting position eases, and clinicians prefer not to inject in this window unless it is severe. People on fluoroquinolone antibiotics or with poorly controlled diabetes are at higher risk of tendon failure under load — start lighter and progress slower, and tell your clinician about the loading you are doing.

What most guides get wrong

"It's inflammation." Acute strain involves some inflammation; the chronic condition that brings most people in does not — the tissue change is degenerative remodelling, not active immune attack Khan et al. 2002. Anti-inflammatories help short-term pain, not the underlying problem.

"Rest will fix it." Complete rest deconditions the tendon further. Pain usually comes back the moment you resume. The mechanical signal a tendon needs to remodel is load, not the absence of load — modified load, applied deliberately, over weeks.

"Get the cortisone shot, you'll be fine." You will be fine at six weeks. The trials say a meaningful fraction of people who took the shot are worse off than the do-nothing arm at one year Smidt et al. 2002 Coombes et al. 2013. One shot for severe pain may earn its place; the second and third should give serious pause Coombes et al. 2010.

"An ultrasound or MRI will tell me how bad it is." They tell you what the tendon looks like, which correlates poorly with how much it hurts. Plenty of bad-looking tendons feel nothing; plenty of normal-looking ones hurt a lot. Imaging is useful for ruling other things out, not for measuring how bad the tendinopathy is.

"Exercise will make it worse." Loading at the wrong dose makes it worse — going back to the racquet at full volume, lifting too heavy too soon. Loading at the right dose is the treatment.

Where the protocol goes wrong in practice

The exercise itself is simple. The reasons it stops working are mostly behavioural.

You went back to full activity the day pain dropped. Pain resolves before the tendon's remodelling does. Phase the racquet, the workbench, the weighted lifts back in over four to eight weeks, not over a weekend.
You wrote the brace into the budget. Counterforce straps unload the tendon while you wear them. They are an adjunct to loading, not a replacement.
You never made the load harder. A protocol that stays at the same resistance bar grade or the same 2 lb dumbbell for two months stops driving the adaptation. Step up when the current load feels routine.
You only treated the elbow. The problem is the chain: a poor backhand technique, a mouse position that pulls the wrist into extension all day, a hammer with too-thin a grip, a shoulder that does not stabilise. Address the upstream load source or the tendon will keep meeting it.
You stacked cortisone injections. The first injection's harm to long-term outcomes is modest; the second and third compound it Coombes et al. 2010. A repeat shot every few months for a year is the worst place to be.
You quit at week three. The Tyler-twist trial showed clear separation from controls between week 4 and week 7 Tyler et al. 2010. Three weeks is the floor for noticing the effect, not the ceiling.

What changes when you do this properly

The first thing that changes is fast. A round of isometric holds calms the pain within a day or two and lets you function without flinching at the doorknob Rio et al. 2015. People who start the loading work in week one usually feel an obvious shift by week three or four.

Weeks 4–8. The trial separation point. The kettle stops being the kettle — you reach for it without the mental flinch you had stopped noticing. Grip strength is recovering, which mostly shows up as the small frustrations that just disappear: opening a jar, holding a phone for a long call, shaking hands without bracing. People around you stop asking what is wrong with your arm because they have stopped seeing you favour it Tyler et al. 2010.

Months 2–3. The sport or the work that the tendon was for comes back in graded volume. Tennis players are hitting again with form intact; carpenters and mechanics are back to a full day without the Thursday flare; new parents are lifting the infant in both arms without the wrist warning. Sleep returns to the affected side. The cortisone shot is the thing you did not take.

The 12-month view. The trials show the people who took the cortisone shot are mostly fine but a meaningful chunk are worse off than the do-nothing arm Smidt et al. 2002 Coombes et al. 2013. The people who did the loading work are in the best position — both on the symptom and on the recurrence number — and they have a protocol they can pull off the shelf the next time the same activity bites Bisset et al. 2006.

For De Quervain's, the timeline compresses. Splint plus modified lifting clears the milder cases in 4–6 weeks; the well-placed cortisone injection clears most of the rest in the following two to four weeks Peters-Veluthamaningal et al. 2009. Postpartum cases that linger past month four usually have a persistent lifting-pattern problem rather than a tendon-that-won't-heal problem.

The kit and the cost

The whole protocol fits in a drawer. A flexible resistance bar in three grades runs roughly $15–30 — start with the lighter colour and work up. A 2 lb (about 1 kg) dumbbell is enough for most wrist exercises in the first month; you may upgrade to 3–5 lb as you progress. An over-the-counter forearm-based thumb spica splint for De Quervain's is $15–25. A counterforce strap if you want one runs the same.

A hand therapist or physiotherapist who sees a lot of tendon work is worth two to four sessions to get the form right, the load right, and the activity-modification audit right — typically $50–200 a session. Coverage varies by system; most need a referral from a primary care clinician, which doubles as the chance to rule out the things on the contraindications list. A corticosteroid injection for De Quervain's, when it is the right call, runs $100–400 depending on whether it is image-guided.

Time is five to ten minutes a day for the loaded exercises plus the audit on the activity that started it — typically a couple of hours over the first week of fiddling with mouse position, grip diameter on tools, or backhand technique, then maintenance.

Who gets which one

The three conditions sort fairly cleanly by who you are and what you do.

Tennis elbow is the most common — somewhere between 1 and 3% of adults each year — with a peak between 40 and 60 and a slight male skew in occupational cohorts. The risk is two to three times higher in people whose work involves repetitive forceful gripping (electricians, plumbers, carpenters, mechanics, hairdressers, butchers, dental work) Shiri et al. 2006 Descatha et al. 2016. Among racquet sports, technique matters more than volume — the late-hitting one-handed backhand is the classic culprit.

Golfer's elbow is roughly one-third as common as tennis elbow at the population level. Same age band. Driven by similar grip-and-twist work, throwing sports, and weight training that loads the forearm flexors heavily (rows, deadlifts with a thick bar, kettlebell swings).

De Quervain's is the one with the demographic signal. It is four to six times more common in women than men, with one peak in the postpartum window (the repeated infant-lift position is the canonical trigger — common enough to have its nickname) and a second peak in middle-aged women in jobs with repetitive thumb work Stahl et al. 2013. The pregnancy and breastfeeding context shapes the treatment — see the contraindications section.

Diabetics, smokers, and people on fluoroquinolone antibiotics have slower tendon healing across the board and should expect a longer protocol and a more cautious load progression.

The other things on the shelf

A handful of treatments live next to loading rehab. Some have a niche; most are weaker than the marketing.

Platelet-rich plasma (PRP) injections. The clinic draws your blood, spins it, and injects the platelet layer into the tendon. The trial evidence is mixed and the meta-analyses do not agree; some show modest benefit over saline at a few months, others find no difference. Reasonable to consider for chronic cases that have not responded to loading, but it is not a first move and not a substitute for the exercises.

Extracorporeal shockwave therapy. Pulsed sound waves to the tendon site. Some positive trials in chronic tennis elbow, some null, and large between-clinic variation in how it is delivered. Similar profile to PRP — a possible adjunct, not a primary.

Dry needling and acupuncture. Short-term pain reduction in several trials; long-term effect not established. Useful if you want a pain bridge while you keep loading.

NSAIDs. Help with the early acute pain. Not a treatment for the underlying tendinopathy and not a replacement for the loading work.

Surgery. Open or arthroscopic release for tennis elbow, and release of the first dorsal compartment for resistant De Quervain's, are real options for the small minority who do not improve after six to twelve months of proper conservative care. They are not first or second-line.

Related things worth knowing about

If this entry is the right neighbourhood for you, three adjacent topics are likely to come up.

Shoulder tendinopathies — the rotator cuff plays by similar load-capacity rules but with anatomy and protocols of its own.
Lower-limb tendinopathies — the Achilles and the patellar tendon have a deeper trial literature that originally set the modern loading approach. Many of the principles travel up.
Nerve entrapments around the elbow and wrist — carpal tunnel, cubital tunnel, radial-tunnel syndrome. These often masquerade as tendinopathy and need different management.
Ergonomic setup — desk posture, mouse and keyboard choice, tool grip diameter, instrument technique. The upstream prevention layer for everything in this entry.
Trigger finger — another stenosing tenosynovitis, sharing some of the De Quervain's logic.

Related in the handbook

Worsens / aggravates

— If a tendon flared up right after an antibiotic course, ask which one — this whole drug class can damage tendons and even rupture them.

Helps

— The loading protocol that actually fixes tennis and golfer's elbow is built on slow eccentric reps.
— Taken with vitamin C before your loading exercises, collagen may give the tendon a little extra raw material to rebuild with.

— Wrist and hand symptoms overlap with carpal tunnel; sort out whether it's tendon or pinched nerve.
— Wrist and forearm tendons often get overloaded at a badly set-up desk; fixing keyboard and mouse position takes the daily strain off while they heal.
— Another tendon that responds to graded loading rather than rest or a quick injection.
— Same story in a different limb: a tendon overloaded faster than it remodels, fixed by graded load, not rest.
— Shoulder cuff pain is the same load-tolerance problem one joint up; the rehab logic carries over.

Substance + claimed effects

The entry covers three common upper-limb overuse tendon disorders: lateral epicondylalgia (tennis elbow, affecting the common extensor origin at the lateral elbow, predominantly extensor carpi radialis brevis), medial epicondylalgia (golfer's elbow, affecting the common flexor-pronator origin), and De Quervain's tenosynovitis (stenosing tenosynovitis of the first dorsal compartment of the wrist — abductor pollicis longus and extensor pollicis brevis). They share a load-capacity pathology, a presentation pattern (load-related point pain plus weakness on the affected grip or pinch), and a treatment principle (relative offload of the aggravating loads plus progressive tendon loading) — which justifies grouping them. Claimed effects of the condition span functional impairment: grip strength loss, reduced ability to perform repetitive manual tasks, sports performance drop, sleep disruption from rest pain in severe cases, and — over months to years if mismanaged — chronicity affecting work and mood. The article projects the literature on what these are, how they recover, and what loading-based rehab actually delivers. Holistic scope per entry.md §1a: every meaningful consequence — function, daily energy and sleep impact from chronic pain, mood, evidence quality — gets a home.

Evidence by addressing question

mechanism

The dominant clinical finding in chronic tendinopathy is not inflammation but tendinosis: disorganised collagen, increased ground substance, neovascularisation, and absent or minimal inflammatory cells on histology Khan et al. 2002. The "-itis" suffix in tendinitis, epicondylitis, and tenosynovitis is therefore semantically misleading for the chronic phase — the BMJ editorial argued for retiring it two decades ago and the sports-medicine literature has largely moved to tendinopathy / epicondylalgia Khan et al. 2002. De Quervain's is a partial exception in that the sheath thickening and stenosis (a mechanical compartment problem) genuinely involves chronic non-inflammatory reactive change of the sheath, not classic acute inflammation either.

Cook and Purdam's continuum model frames tendinopathy as load-capacity mismatch playing out across three overlapping stages: reactive tendinopathy (short-term cell and matrix response to acute overload, reversible with relative rest), tendon disrepair (greater matrix breakdown, partial reversibility), and degenerative tendinopathy (focal cell death and matrix disorganisation, limited reversibility but capable of becoming asymptomatic) Cook & Purdam 2009. The clinical implication: the treatment is not anti-inflammatory; it is graded loading that pushes capacity back above demand. Pain is not directly proportional to tissue damage on imaging — many asymptomatic tendons show degenerative change and many painful tendons look unremarkable on ultrasound.

Mechanistically, repetitive eccentric loading of the wrist extensors (lateral elbow) or flexor-pronators (medial elbow) at the muscle-tendon junction generates focal microtearing at the bone-tendon interface; recovery requires both an offload from the provoking task and a tendon-specific stimulus that drives collagen remodelling. For De Quervain's the proximate mechanism is friction and impingement of the first dorsal compartment tendons through a thickened retinaculum, with repetitive thumb abduction and ulnar wrist deviation as the canonical aggravators (the postpartum thumb-and-wrist position of holding an infant is the classic example) Stahl et al. 2013.

evidence

Three large pragmatic RCTs anchor the lateral-epicondylalgia evidence base, and they tell a consistent story about natural history vs. corticosteroid injection vs. exercise.

Smidt et al. 2002 in The Lancet randomised 185 patients to corticosteroid injection, physiotherapy (manual therapy + exercise), or wait-and-see. At 6 weeks corticosteroid was best (92% success); by 52 weeks it was the worst (69% success vs. 91% wait-and-see, 91% physiotherapy), with a recurrence rate of 72% in the steroid arm Smidt et al. 2002. Bisset et al. 2006 in BMJ replicated the pattern with 198 patients across the same three arms: corticosteroid injection gave short-term wins but worse outcomes at 52 weeks compared to either physiotherapy or wait-and-see, with recurrence rates above 70% Bisset et al. 2006. Coombes et al. 2013 in JAMA ran a 2×2 factorial of corticosteroid (vs. saline) and physiotherapy (vs. none); at one year, corticosteroid had significantly worse complete recovery / no-recurrence rate (54%) than placebo injection (78%), and adding physiotherapy did not rescue the steroid arm Coombes et al. 2013. The systematic review of injection therapies across tendinopathies confirmed the pattern: short-term benefit, worse intermediate and long-term outcomes versus other interventions or no treatment Coombes et al. 2010.

For loading-based rehabilitation in lateral epicondylalgia, the Tyler eccentric protocol (FlexBar / "Tyler twist") added to standard physiotherapy beat standard physiotherapy alone — large effect sizes on pain and DASH score over 7 weeks in a small RCT Tyler et al. 2010. Systematic review of eccentric exercise for lateral epicondylitis (12 studies) supported a positive but heterogeneous effect, with eccentric or eccentric-concentric protocols outperforming controls on pain and function Cullinane et al. 2014. The 2019 international consensus on tendinopathy (ICON 2019) endorsed progressive loading as the cornerstone across tendinopathies Vicenzino et al. 2020. Isometric loading produces immediate analgesia in patellar tendinopathy via cortical inhibition reduction — the mechanism likely generalises to upper-limb tendinopathies and is the rationale for using isometric holds when eccentrics are too provocative Rio et al. 2015.

For De Quervain's, the Cochrane review of corticosteroid injection (1 RCT, 3 quasi-RCTs, n=126) found injection more effective than splinting or NSAIDs at 1–6 weeks, with cure rates near 83% after injection compared to 14% with splinting alone — De Quervain's is the one upper-limb tendinopathy where corticosteroid retains a defensible role, particularly outside pregnancy/postpartum Peters-Veluthamaningal et al. 2009. Surgical release of the first dorsal compartment is reserved for failure of conservative care after months.

The meta-analysis of nonsurgical treatment for lateral epicondylitis across heterogeneous interventions suggested that, at one year, most active treatments converge with watchful waiting in mean outcomes — the natural history is favourable for the majority and intervention's main job is to shorten time-to-resolution and reduce flare severity, not to determine whether the tendon eventually heals Sayegh et al. 2015. Roughly 80–90% of lateral epicondylalgia resolves within 12–24 months regardless of treatment Cutts et al. 2020.

protocol

For lateral epicondylalgia, the canonical loading protocol is the Tyler twist using a flexible resistance bar (FlexBar): wrist extension on the affected side, with the contralateral hand bending the bar further, then a slow (3–4 s) eccentric lowering as the affected wrist allows the bar to untwist — three sets of 15 reps, once daily, 7 weeks Tyler et al. 2010. Heavy-slow-resistance with dumbbell wrist extension (slow 3–4 s eccentric, 3 sets × 10–15 reps, 3×/week) is a functionally equivalent substitute drawn from lower-limb HSR protocols. For medial epicondylalgia, the symmetric protocol replaces wrist extension with wrist flexion / pronation eccentrics; the FlexBar "reverse" twist applies.

Isometric wrist extension or flexion holds (45 s × 5 sets, 70% MVIC) are useful when full-ROM eccentrics provoke too much pain — they reduce pain immediately by 24–48 hours Rio et al. 2015 and let loading start at a lower irritability tier. The general loading principle across tendinopathies (per ICON 2019): start at the highest load the tendon tolerates the next morning at ≤4/10 pain, progress weekly, do not push through escalating pain over consecutive sessions Vicenzino et al. 2020.

For De Quervain's: thumb spica splint (forearm-based or short opponens), avoidance of the lift-with-thumb-abducted maneuver (the babylift position), short course of NSAID for severe pain, and corticosteroid injection if splinting plus activity modification fails in 4–6 weeks Peters-Veluthamaningal et al. 2009. Loading rehab for De Quervain's is less codified than for the elbow tendinopathies because the proximate problem is sheath stenosis; gentle thumb extension/abduction strengthening returns after symptoms subside.

Across all three: identify and reduce (not eliminate) the provoking load — backhand grip on a tennis racquet, mouse posture, hammer grip, infant lift — for a 4–8 week recovery window, then gradually reintroduce. Counterforce bracing (elbow strap distal to the epicondyle) provides modest short-term pain reduction during loading; not a stand-alone treatment.

contraindications

The progressive-loading approach assumes the diagnosis is correct. Red flags that prompt referral rather than self-managed loading: true rupture (sudden tearing sensation, palpable defect, loss of active function), radicular pain from C6/C7 referral mimicking lateral elbow pain, radial nerve entrapment (radial tunnel syndrome — pain 4 cm distal to the lateral epicondyle, exacerbated by middle finger resisted extension), posterior interosseous nerve palsy, intersection syndrome (more proximal forearm pain than De Quervain's, with crepitus), and any constitutional features (fever, night pain unrelated to position, systemic weight loss) that suggest a non-mechanical cause. Inflammatory arthropathies (rheumatoid, psoriatic) and crystal disease produce tendon-area pain and respond very differently. In pregnancy and breastfeeding, corticosteroid injection for De Quervain's is typically deferred — the condition often resolves with cessation of breastfeeding and the postural triggers, and injection in this window is reserved for severe cases.

misconceptions

The label tendinitis still carries the implication of acute inflammation and points the reader toward anti-inflammatory treatment. The histology says otherwise for the chronic presentation that brings most people in Khan et al. 2002. A second misconception: rest cures it. Complete rest deconditions the tendon and unloads the very stimulus needed to remodel; pain typically returns when activity resumes. A third: the cortisone shot fixes it. Short-term it does; at one year it makes things worse in the elbow tendinopathies Smidt et al. 2002 Coombes et al. 2013. A fourth: an ultrasound or MRI confirms how bad it is. Imaging findings correlate poorly with pain — many degenerative tendons are asymptomatic and many painful ones look unremarkable. A fifth: exercise will make it worse. Loaded exercise that respects irritability is the treatment; the threshold is morning-after pain ≤4/10 and non-escalating Vicenzino et al. 2020.

stakes

Untreated lateral epicondylalgia has a favourable natural history at the population level — roughly 80–90% resolve within 1–2 years Sayegh et al. 2015 Cutts et al. 2020. The individual reader's experience is the median case: pain that is one week tolerable, one week limiting, sleep disruption when the rest pain phase hits, grip strength deficit of 20–40% during symptomatic months that affects daily tasks (kettles, jars, doorknobs, holding a phone), reduced participation in racquet sports or workshop work, and — with repeated corticosteroid injections — meaningfully worse long-term outcomes including elevated recurrence risk Coombes et al. 2010. The minority who become chronic (10–20%) face symptoms beyond two years, work-time loss, and in occupational cases career-change considerations. Population epidemiology: 1–3% annual prevalence for lateral epicondylalgia, peaking 40–60 years, higher in workers exposed to repetitive forceful gripping and awkward wrist postures Shiri et al. 2006 Descatha et al. 2016. For De Quervain's, postpartum cases that don't get managed often persist through the first year of infant-lifting; for working-age cases with continued occupational aggravation, persistence is common.

payoff

Progressive loading with the aggravating activities tapered shortens time-to-resolution and reduces flare intensity compared to repeat steroid injections Smidt et al. 2002 Bisset et al. 2006 Coombes et al. 2013. Time-course of felt benefit: isometric holds give pain reduction within 24–48 hours of the first session Rio et al. 2015; the Tyler eccentric protocol shows separation from controls by 4–7 weeks Tyler et al. 2010; grip strength normalisation lags pain resolution by weeks to months; full return to provoking sport / heavy occupational work typically takes 3–6 months for the elbow conditions. For De Quervain's, injection plus splinting often resolves symptoms in 2–6 weeks; for postpartum cases, weaning of the infant-lift posture is the often-overlooked second half of the protocol.

failure-modes

The most common rehab failures: (1) pain catastrophising into total avoidance — the tendon needs loading; complete rest delays recovery and predicts chronicity. (2) Returning to provoking activity at full volume because pain is gone — pain resolves before tendon remodelling completes; a graded return over 4–8 weeks reduces recurrence. (3) Stacking corticosteroid injections — second and third injections compound the long-term harm seen in the trials Coombes et al. 2010. (4) Using a brace as a substitute for loading — counterforce straps are an adjunct, not a treatment. (5) Insufficient load progression — the protocol that doesn't get heavier doesn't drive remodelling. (6) Treating one elbow without addressing the kinematic chain — backhand technique, mouse posture, grip size on tools, and shoulder/scapular control all upstream of the elbow.

practicalities

Equipment is cheap: a FlexBar in three resistance grades runs roughly $15–30; a 2 lb (1 kg) dumbbell is sufficient for most wrist HSR; a forearm-based thumb spica splint is $15–25 over the counter. Hand therapist or physiotherapist visits cost $50–200 per session, typically 4–8 sessions over a course. Insurance coverage varies; a written diagnosis from a primary care clinician helps in most systems. Time commitment is 5–10 minutes daily for the loading exercises plus the time to identify and modify the provoking activity. Corticosteroid injection for De Quervain's runs $100–400 depending on imaging guidance and system.

out-of-scope

Adjacent topics not covered: rotator cuff tendinopathy (similar load-capacity principle, different anatomy and different protocol), Achilles and patellar tendinopathy (the lower-limb tendinopathy literature is older and richer; many principles transfer up but the protocols differ), cubital tunnel and carpal tunnel (nerve entrapments often confused with tendinopathies), trigger finger (another stenosing tenosynovitis), RSI / repetitive strain umbrella (a working diagnosis covering many specific conditions including these), and ergonomic setup for desk and trades work (the upstream prevention layer).

The credibility range

Optimist case. The pathology is well-characterised (Cook-Purdam continuum), the failure of anti-inflammatory framing is established across multiple high-quality RCTs (Smidt, Bisset, Coombes), the loading-based rehab principle is endorsed by international consensus (ICON 2019), and the protocols (Tyler twist, isometric holds, heavy-slow-resistance) have direct or by-analogy RCT support. Natural history is favourable for most. A reader following the literature-aligned protocol can expect meaningful pain reduction within weeks and high probability of resolution within 6–12 months — substantially better than the do-nothing or repeated-cortisone trajectory. For De Quervain's specifically, a single well-placed corticosteroid injection has high cure rates documented in a Cochrane review.

Skeptic case. The natural history is so favourable that distinguishing intervention effect from regression to the mean is hard. The headline finding — that steroid injection worsens long-term outcomes — is robust in the elbow but rests on three RCTs; the loading-protocol superiority over wait-and-see is less clearly established at one year for the typical patient (Sayegh meta-analysis). Eccentric-loading RCTs are small, heterogeneous, and outcome-measure-inconsistent; effect sizes shrink in pooled analyses. Many "tennis elbow" diagnoses are actually radial tunnel syndrome or referred cervical pain, and reported success of any treatment is partly a sampling artefact. The transfer of patellar-tendon isometric analgesia to upper-limb conditions is plausible-but-unproven. PRP and other adjunct injections have noisy evidence with publication bias.

Author's call. The progressive-loading + cortisone-skepticism stance is the right one for elbow tendinopathies; the evidence is strong enough for confident recommendation. The loading protocol details (eccentric vs. heavy-slow-resistance vs. isometric) are second-order and any of them beat passive treatment. For De Quervain's, splinting plus activity modification first; corticosteroid injection second-line and high-yield outside the pregnancy/breastfeeding window. Evidence: 4 (multiple consistent RCTs for the elbow story; Cochrane for the De Quervain's story; one good RCT-equivalent for the loading protocol). Controversy: 2 (genuine but minor debate on optimal loading parameters, on PRP/dry needling, on the role of isometric loading; no foundational disagreement).

Stakeholder + incentive map

Commercial: FlexBar / Theraband (resistance-bar makers), brace and splint manufacturers, PRP injection providers (clinics with margin on autologous blood products), shockwave therapy providers (capital equipment, paid sessions). PRP and shockwave have weak evidence but strong commercial push.
Professional: Sports-medicine physicians and physiotherapists / hand therapists are aligned with the loading-based protocol (well-represented in the ICON consensus). Primary care still defaults to NSAIDs ± steroid injection in many systems; orthopaedic surgeons see the chronic-resistant cases and may over-index on surgical release for De Quervain's.
Community: Racquet sports communities (especially recreational tennis, pickleball) and trades communities (electricians, carpenters, mechanics) generate a lot of self-reported anecdata about counterforce straps, technique fixes, and equipment changes. Postpartum forums carry significant De Quervain's signal. Quality of community advice is mixed; the rest-and-NSAID frame is widely repeated.
Counter / skeptic: Cochrane and the high-quality elbow RCTs are the counter-pressure on the cortisone-first reflex. Hand-therapy guidelines (American Society of Hand Therapists) and AAOS clinical practice guidelines have moved cautiously.

Population variability

Lateral epicondylalgia peaks 40–60 years, slight male predominance in occupational cohorts; 1–3% annual prevalence in adults; risk elevated 2–3× by repetitive forceful gripping, awkward wrist postures, and high-force occupational work Shiri et al. 2006 Descatha et al. 2016. Medial epicondylalgia is roughly one-third the prevalence of lateral, similar age band, similar work exposures. De Quervain's has a strong female predominance (4–6× more common in women), bimodal peak with one peak postpartum (the "mommy thumb" pattern from repeated infant lift) and a second peak in middle-aged women, and meaningful occupational risk in jobs with repetitive thumb abduction and ulnar wrist deviation Stahl et al. 2013. Athlete subpopulations: tennis (backhand mechanics drive lateral), golf (medial), throwing sports (medial), bowling and racquetball mixed. Diabetics and patients on fluoroquinolones have elevated tendon-failure risk that modifies the loading approach. Smokers have slower tendon healing. Hypothyroidism and obesity are associated with elevated prevalence across tendinopathies.

Knowledge gaps

Where the literature is thin: medial epicondylalgia has far fewer dedicated RCTs than lateral; most protocols are adapted by symmetry rather than tested. Optimal loading parameters (eccentric vs. heavy-slow-resistance vs. isometric, frequency, progression) are not head-to-head established in the upper limb. PRP, autologous blood, dry needling, shockwave all have RCT data of mixed quality; meta-analyses do not converge. Long-term outcomes of De Quervain's beyond 1 year, and recurrence rates after a single steroid injection, are under-reported. The cortical / pain-modulation mechanism of isometric loading shown in patellar tendinopathy Rio et al. 2015 has not been directly demonstrated for the elbow or wrist tendons. Ergonomic intervention evidence is observational and confounded; controlled trials are scarce. What would change the call: a high-quality multi-arm RCT comparing isometric, eccentric, and HSR loading in lateral epicondylalgia with 2-year follow-up; head-to-head PRP-vs-placebo with adequate sample for the small effect; a properly designed RCT for medial epicondylalgia.

Scope vs. brief. Brief named tennis elbow, golfer's elbow, and De Quervain's; effects on grip, daily function, sport and work performance; recovery via progressive loading. All five named consequences are covered end-to-end. The three conditions are grouped because they share the load-capacity pathology, the presentation pattern, and the loading-based treatment principle — splitting them into three entries would triple-state the mechanism and the misconceptions paragraph. De Quervain's gets a flagged exception on cortisone in the evidence + protocol sections so the reader does not over-generalise the elbow finding.

Hard scoping calls. Rotator cuff tendinopathy excluded — same family but a substantively different protocol and clinical workup; deserves its own entry. Lower-limb tendinopathies excluded — the Alfredson Achilles literature and the patellar HSR work are the historical roots of the loading principle but the protocols differ. Trigger finger and nerve entrapments named in out-of-scope but not covered. RSI / repetitive strain umbrella avoided as a frame — too imprecise.

Rating difficulties. Action sat between respond and do. Chose respond because the entry's value is overwhelmingly to the reader who already has the symptom or is at the threshold; the prevention layer is ergonomic and lives in another entry. Cadence: course — the loading protocol is bounded (~8–12 weeks). Evidence set at 4 not 5: the lateral epicondylalgia trial stack is genuinely 5-tier; the medial epicondylalgia literature is thinner and the protocol is by-symmetry; pooling the two pulls the holistic rating down a notch. Controversy at 2: the cortisone-skepticism call is robust; the loading-parameter debates (eccentric vs HSR vs isometric, PRP, shockwave) are real but second-order.

Future-link candidates. rotator-cuff-tendinopathy, lower-limb-tendinopathies (or per-tendon equivalents), carpal-tunnel-syndrome, cubital-tunnel-syndrome, trigger-finger, desk-ergonomics, tool-grip-and-hand-position. When any of these land, wire out-of-scope to them.

Separate-entry candidates surfaced during the write. A standalone De Quervain's tenosynovitis entry could be justified given the distinct demographic (postpartum), the cortisone-yes story, and the splinting protocol — but the editorial trade is favourable here, since grouping it lets the reader see the cortisone-question framed as "not always wrong, but mostly wrong in this family". A Cortisone injection for musculoskeletal pain entry would also be high-value (similar long-term-loss pattern in plantar fascia, shoulder).

Citations used vs. dossier superset. Dossier carries 16 refs; article uses 12. Refs not surfaced in the article body: Cook & Purdam 2009 sits behind the mechanism narrative but is not explicitly cited in the article text (only in the mechanism dossier para and one article reference); Cullinane et al. 2014 is dossier-only. This is the intended superset relationship.

One judgment call worth flagging. The audience-section mention of fluoroquinolone-associated tendon risk and diabetes-slowed healing is a real and well-described modifier but is not central to the entry. Kept it brief; did not promote to its own callout because the typical reader is neither, and over-emphasising would dilute the loading-protocol punchline.