RightRep EN · PT Open the app
BLOG

Can You Train With Shoulder Impingement? The Subacromial Audit: Where the Diagnosis That Named a Generation of Painful Shoulders Failed Placebo Surgery Twice — Oxford's CSAW Trial Watching Decompression Score 32.7 Against Sham Arthroscopy's 34.2, Finland's FIMPACT Finding No Benefit Over Placebo or Exercise at Two Years and Again at Five, Scotland's Surgeons Answering by Cutting the Operation 29% in Four Years; Where a Third of Shoulders That Never Hurt Carry the Same Subacromial Spurs the Scan Gets Blamed For and the Field's 2025 Position Retires the Label to 'Rotator Cuff-Related Shoulder Pain'; Where the Gym's Own Census Associates the File With Upright Rows and Lateral Raises Above 90° and With Missing External-Rotator Strength, and Where Holmgren's Twelve-Week Specific-Exercise Trial Gained Constant Points 24 to 9 and Cut the Choice of Surgery From 63% to 20%: Train — the Pain Is Real, the Villain Is Not, and Progressive Loading Is the Only Door With Numbers Behind It

By RightRep AI Assistant · · 21 min read

There is a fence at the top of this post and it outranks everything below it: shoulder pain that arrived with a fall, a pop, sudden weakness, an arm that will not rise, or night pain that wakes you is the tear-and-trauma door — a clinician first, and if the ultrasound or MRI report says full-thickness tear, that is the rotator-cuff tear audit's file, not this one. The second fence: a shoulder that is stiffening — rotation and behind-the-back reach shrinking week over week, not just hurting — is the frozen shoulder file, which runs on its own clock and answers to different rules. The third: pain sitting exactly where the collarbone ends on top of the shoulder, worse reaching across the body, belongs to the AC joint — a different joint with its own mechanism file. The fourth: symptoms running below the elbow, or neck positions that light the shoulder up, are the neck-pain file's referral question wearing a shoulder costume. The fifth: fever, feeling unwell, a cancer history, unexplained weight loss, or weakness that progresses week over week is the today door — clinician now, not a program. With the fences respected, the question underneath is the most common shoulder file in the gym, and this library's shoulder corner had every piece of it except the center: the label on the pain ("impingement," "bursitis," "tendonitis") versus the operation that label was built to justify. The evidence does something rare here — it dissolves the diagnosis itself, fails the surgery against placebo surgery twice, and hands the whole file to progressive loading with the biggest numbers in the corner [1][5][6][11].

The short answer: train — everything the shoulder doesn't need from day one, while the shoulder itself loads specifically and progressively

Strip the file to its spine. The label is dissolving: the field's 2025 position paper argues that neither clinical tests nor imaging can definitively implicate the blamed structures, and clinicians worldwide now choose "rotator cuff-related shoulder pain" over "impingement" — the pain is real, the villain is not [1]. The scan is not the verdict: among 100 volunteers whose shoulders never hurt, MR imaging found subacromial spurs in a third and AC-joint arthritis in three quarters — the "impingement findings" are background noise in pain-free people, not a smoking gun [2]. The operation failed placebo surgery twice: Oxford's CSAW trial randomized 313 patients to decompression, sham arthroscopy, or nothing — and the real operation scored below the sham at six months (32.7 vs 34.2 points, difference −1.3, p=0.31) [5]; Finland's FIMPACT trial found no clinically relevant benefit over placebo surgery at two years, and none at five [6][7]. The gym owns its own census: impingement characteristics cluster in weight-trainers, associate with upright rows and lateral raises above 90°, and sit inversely with external-rotator strength — the file has a barbell-shaped risk profile, which means it has a barbell-shaped program [3]. And the strongest treatment door in the whole corner is loading: Holmgren's twelve-week specific-exercise trial — eccentric rotator-cuff work plus scapular-stabilizer training — improved Constant scores 24 points to the control's 9, and cut the share of patients choosing surgery from 63% to 20% [11]. Train legs, trunk, the healthy arm, conditioning — all of it — from day one; keep the pressing arc below the painful range and rebuild the rotator cuff and scapular stabilizers with the patience the trials used; and let twelve weeks of honest, logged loading — not an operation — write the verdict [5][6][11].

Who is asking: the most common shoulder file in the gym

This is not a rare file hiding in a specialist's corridor; it is the default setting of the painful shoulder. "Impingement syndrome" spent four decades as the most common shoulder diagnosis in sports medicine — the label under which most painful-shoulder-in-the-gym stories were filed — and the weight room turned out to have its own epidemiology. Kolber's research group did the study this library needed: seventy-seven men, mean age 28 — forty-six who weight-trained at least twice a week and thirty-one controls who never lifted — examined with the two-test cluster (painful arc plus Hawkins-Kennedy) that carries a positive likelihood ratio of 5.0 against diagnostic gold standards [3]. The weight-training group carried significantly more impingement characteristics than the controls (p<0.001), and inside the training questionnaires the associations sharpened: lateral deltoid raises and upright rows performed above 90° associated with impingement characteristics (p≤0.004), while external-rotator strengthening carried a significant inverse association — the exercise that protects against the file is the one most programs skip [3]. The follow-up study quantified what the impinged lifter actually looks like: among fifty-five recreational weight-trainers, the twenty-four with impingement signs showed decreased internal- and external-rotation range (p≤0.016), decreased bodyweight-adjusted external-rotator and lower-trapezius strength (p≤0.02), and worse agonist-to-antagonist strength ratios (p≤0.004) than the thirty-one without [4]. That is the census in one sentence: the lifter with this file is stronger in front, tighter in rotation, weaker behind, and trains the exact exercises the associations flag — which is also why the file answers to training [3][4].

The naming door: the diagnosis that is being retired

Every audit family has a naming file — the tendons had tendinopathy not tendinitis, the groin had Doha — and this corner's is a retirement notice. The word "impingement" was built as a mechanical story: something under the acromion getting pinched, with the surgery named after the story (decompress the space, free the pinched tissue). The field has spent a decade taking that story apart, and the 2025 position piece says the quiet part in print: in the absence of diagnostic accuracy for clinical and imaging tests, specific shoulder structures cannot be implicated as the definitive source of symptoms — the rotator cuff tendons are not the proven sole villain, terms like "tendinopathy" and "tendinitis" overclaim, and in a global survey of clinicians the label that won "overwhelming first choice" was deliberately structure-neutral: rotator cuff-related shoulder pain [1]. Two consequences travel with the retirement. First, the pain is not being dismissed — the new name relocates it from a pinched structure (which surgery claims to fix) to a load-tolerance problem in the rotator cuff and scapular system (which loading actually treats) [1]. Second, the search for the one guilty structure on the scan was misdirected from the start — which is the next door [1][2].

The scan door: the findings live in shoulders that never hurt

If the diagnosis were really a pinched structure visible on imaging, the findings should track the pain. They do not. The cleanest demonstration is older than the debate itself: one hundred asymptomatic volunteers, aged 19 to 88, MR-imaged shoulder by shoulder — and among shoulders that had never hurt, characteristic AC-joint osteoarthrosis sat in three quarters, subacromial spurs in one third, joint fluid in nearly all, with bursal and peribursal changes simply paralleling whatever the cuff tendons looked like [2]. The authors' conclusion is the scan door in one line: a high prevalence of MR-evident bone and peritendinous abnormalities among asymptomatic individuals — the finding is not a verdict [2]. Put that together with the placebo-surgery door below and the whole mechanical story wobbles: the spur the scan blames is common in pain-free shoulders, and shaving the underside of the acromium — the operation built on the spur — does not beat placebo surgery [2][5][6]. The honest translation for the lifter holding a report that says "bursitis" or "spur": the report describes the neighborhood, not the culprit, and the treatment decision never needed the scan in the first place [1][2].

The placebo-surgery door: the operation that failed the hardest test twice

Medicine has few tests harder than placebo surgery — open the shoulder, do nothing therapeutic, close — and this diagnosis has now faced it twice, in the two trials that reorganized the entire corner. Oxford's CSAW trial (Lancet, 32 UK hospitals, 51 surgeons) took 313 patients with at least three months of subacromial pain, intact cuff tendons, who had already failed exercise therapy and at least one steroid injection — the exact population surgeons decompress — and randomized them three ways: arthroscopic subacromial decompression, investigational arthroscopy only (the placebo: the operation without the essential bone-and-soft-tissue removal), or no treatment [5]. At six months, the primary Oxford Shoulder Score read 32.7 for decompression, 34.2 for the sham, 29.4 for nothing — the real operation a statistically indistinguishable 1.3 points behind the sham (95% CI −3.9 to 1.3, p=0.31); both surgical arms did beat no-treatment (+2.8, p=0.019; +4.2, p=0.0014), under the trial's own 4.5-point clinical-importance bar; six complications, every one of them a frozen shoulder [5]. Finland's FIMPACT trial (BMJ, 210 patients, three hospitals) ran the same design with an exercise-therapy arm: at 24 months, decompression versus diagnostic arthroscopy differed by −4.6 VAS points at rest (p=0.18) and −9.0 on activity (p=0.054) — nowhere near the pre-set 15-point threshold [6]. And at five years, with 83% of the trial still followed, the differences were −2.0 at rest (p=0.56) and −8.0 on activity (p=0.093) against the sham, and 1.0 (p=0.77) and −3.9 (p=0.40) against exercise — the authors' own conclusion: decompression "provided no benefit over diagnostic arthroscopy (or exercise therapy) at 5 years" [7]. One operation, two placebo-controlled trials, zero wins.

The surgery-honesty door: what practice did next, and the small print told straight

This audit family does not do one-sided files, so here is the surgery file in full. The meta-analysis that predated the placebo trials pooled four RCTs of conservative versus surgical care and found no outcome differences in any of them, irrespective of methodological quality [8]. The Finnish trial with the longest arc — 140 patients, acromioplasty-plus-exercise versus exercise alone — reached a mean twelve years of follow-up: both arms improved significantly over baseline pain, with no significant between-group difference on any measure, and the detail that belongs in every file: the operated patients carried a stronger belief in recovery, which the authors read as surgical context and placebo — the operation's most durable product was conviction [9]. That trial's final review is titled bluntly: arthroscopic decompression not recommended in the treatment of rotator cuff tendinopathy [9]. Practice then moved: Scotland's surgical census counted decompressions falling from 2,217 to 1,568 in four years (−29%, standardized rate 41.6 to 28.9 per 100,000), with the steepest drop in 2017–2018 — the years the placebo trials published [10]. And the small print, honestly: both CSAW surgical arms did outperform doing nothing (a few points on a 48-point scale, under the clinical bar), and FIMPACT's secondary comparison did favor decompression over exercise statistically at two years (−7.5 rest, −12.0 activity, both p<0.01) — but under the 15-point MID, unblinded, and biased toward surgery by selective exclusions, as the authors themselves flagged [5][6]. The surgery file's honest summary: an operation that beats nothing slightly, beats placebo never, and was already being abandoned by the systems that measured it [5][6][7][8][9][10].

The exercise door: the only intervention with big numbers

Against that backdrop, the exercise trials are the corner's counterweight. The strongest is Holmgren's: 102 patients whose subacromial impingement had persisted over six months after failing earlier conservative treatment — the queue-for-surgery population — randomized to twelve weeks of a specific exercise strategy (eccentric rotator-cuff strengthening plus concentric/eccentric scapular-stabilizer work, with manual mobilization, five to six guided sessions and daily home work) versus unspecific neck-and-shoulder movement exercises [11]. The specific group improved its Constant-Murley score 24 points against 9 (between-group difference 15, 95% CI 8.5–20.6); 69% versus 24% reported a successful outcome (OR 7.6); and the decision that matters most — 20% versus 63% chose surgery (OR 7.7) [11]. Around that spine sit the matches and the origins: Haahr's trial randomized 90 patients (aged 18–55, symptoms six months to three years) to arthroscopic decompression or graded physiotherapy training, and at twelve months the Constant score had climbed 34.8→57.0 with training versus 33.7→52.7 with surgery — non-significant; "surgical treatment was not superior to physiotherapy with training" [12]. Brox's 1993 origin trial randomized 125 patients three ways — acromioplasty, supervised exercise for three to six months, or twelve sessions of detuned (placebo) soft laser — and found surgery and supervised exercise "significantly, and equally, improved" rotator cuff disease compared with placebo, at a difference of 4.0 Neer points (95% CI −2 to 11) between them and a cost of £720 surgical against £390 exercise [13]. Read the three together and the exercise door's sentence: matched the operation in every honest comparison, beat placebo where the operation couldn't, and produced the only large effect sizes in the corner — with the dosage written on the label: specific, progressive, supervised, twelve weeks [11][12][13].

The mechanism door: capacity, not angles

What actually changes when loading works? The consensus statement on the scapula supplies the frame: scapular dyskinesis — altered position and motion of the shoulder blade — is present in a high percentage of most shoulder injuries, impingement symptoms are particularly affected by it, but its exact causal role is not clearly defined; it is best viewed as a potential impairment to shoulder function, one a reliable clinical evaluation can see and rehabilitation can address [14]. The gym version of that impairment is Kolber's deficit profile: less rotation range, less external-rotator and lower-trapezius strength, worse ratios — every item trainable [4]. The Brazilian files then deliver the mechanism door's twist. Camargo's São Carlos group (UFSCar) randomized 46 patients to four weeks of stretching-and-strengthening with or without manual therapy: pain, function and pressure-pain thresholds improved similarly in both groups while scapular kinematics changed only in small, clinically irrelevant amounts — the authors' conclusion being that the improvements are "not likely explained by changes in scapular kinematics" [15]. Haik's thoracic-manipulation RCT from the same group sharpened it: 61 patients, two sessions over a week, and function scores improved in both arms while the manipulation's own measurable effect was a fraction of a pain point [16]. The mechanism door's honest sentence: the pain responds to load capacity, not to x-ray angles, crack techniques or shoulder-blade choreography — train the cuff's strength, the stabilizers' endurance and the range, and skip the search for the perfect scapular position [14][15][16].

The training-through door: what trains, what modifies, what stops

With the fences cleared, the program writes itself along the family's rules — and this file's program is unusually concrete because the census named the dials. What trains from day one: everything the shoulder doesn't need — legs, trunk, the healthy arm (double duty, with the crossover-strength honesty this library has already filed), conditioning; the painful shoulder's pain-free arcs train too, because the file is a load-tolerance problem and complete rest is the one prescription no trial supports [11][12][13]. What modifies: the two exercises the census associated with the file — lateral raises and upright rows above 90° — get re-gated below the painful range or swapped (wide-grip upright rows for face pulls or strict rows; the bench-press audit's grip-width and elbow mechanics apply to pressing); overhead work runs in the arc that stays comfortable and reclaims depth as tolerance grows, dosed by the 24-hour traffic light — reactivity that settles by morning keeps the ladder climbing, reactivity that compounds steps it down [3]. What graduates from afterthought to mainline: exactly what the deficit profile says is missing — external-rotator work and lower-trapezius work, progressed with real load like any other lift, plus rotation range; the protective association in the census was with external-rotator strengthening, not token bandwork [3][4]. What stops: the fences — trauma, the tearing weakness, the stiffening shoulder, the AC-joint story, the neck-shaped symptoms, the red flags; plus the twelve-week honesty: a shoulder that has run an honest, logged, specific program and is no better belongs back with a clinician, not on a harder version of the same program [5][11]. Underneath it all, the family's standard loading logic applies — the tendonitis file's isometrics-to-eccentrics ladder, a real warm-up, progressive overload with deload weeks, and a training log as the dose ledger.

So what do you actually do? The practical file

Six rules, every one sourced from the files above:

  1. Clear the fences before you program the shoulder. A fall, pop, sudden weakness, an arm that won't rise or night pain that wakes you is the tear door — clinician first, and a full-thickness tear on the report is the rotator-cuff tear file; a stiffening shoulder is the frozen shoulder file; collarbone-end pain is the AC file; below-elbow symptoms or neck-triggered pain is the neck file; fever, weight loss or progressive weakness is clinician-today [5].
  2. Do not buy the villain. The label is retiring because no test pins the guilty structure, and the scan findings live in a third of pain-free shoulders — "bursitis" and "spur" on a report describe the neighborhood, not the culprit, and nothing about them predicts who needs surgery [1][2].
  3. Run the twelve weeks the trials ran. Specific, progressive, supervised-where-possible loading — eccentric cuff work, scapular-stabilizer strength, rotation range — five to six guided sessions with daily home work in Holmgren's dosing; that is the program that gained 24 Constant points against 9 and cut surgery choice 63%→20% [11].
  4. Re-gate the two flagged exercises; program the protective one. Lateral raises and upright rows above 90° carry the association; keep them below the painful arc or swap them; external-rotator and lower-trapezius strengthening carries the inverse association — make it a tracked, progressively overloaded mainline, not a bandwork ritual [3][4].
  5. Do not confuse the operation with the answer. Decompression lost to sham arthroscopy at six months, tied placebo at two and five years, matched exercise for twelve years, and Scotland cut it 29% once the trials published; if surgery is offered after a failed specific program, ask which file you are in — the CSAW population had already failed exercise and injections [5][6][7][9][10].
  6. Keep everything else training and log the dose. Legs, trunk, the healthy arm, conditioning and the pain-free arcs — from day one; the verdict is the twelve-week trend in the log, morning reactivity beside every load, the arc depth beside every press [11].

How RightRep fits into this

The subacromial file is, at its core, a twelve-week dose-tracking question — and it is unanswerable without a record:

  • The dose memory. What loaded the shoulder — the upright-row volume, the press arc, the week the lateral raises went above 90° — is visible in a logged history before it is visible in the shoulder [3].
  • The everything-else program. Rule six is a full-body program that skips nothing but the fence — counted, scored sessions keep legs, trunk and the healthy arm progressing through the rebuilding weeks [11].
  • The arc and ratio ledger. The file's grading currency is what the shoulder tolerates at which depth — dated notes beside every session, morning reactivity beside every load, the trend the twelve-week verdict reads [3][4].
  • Form, rep by rep. Camera-scored reps keep the rebuilding ladder honest — the elbow path and pressing mechanics the program asked for, at the tempo it prescribed, no silent drift past what the budget allows.

The bottom line

  • The label: retiring — neither clinical tests nor imaging can definitively implicate the blamed structures, and clinicians' overwhelming first choice in a global survey is now "rotator cuff-related shoulder pain" [1].
  • The scan: 100 pain-free shoulders — subacromial spurs in a third, AC-joint arthritis in three quarters; the "impingement findings" are background prevalence, not a verdict [2].
  • The census: 77 men — lifters carried more impingement signs than controls (p<0.001); lateral raises and upright rows above 90° associated (p≤0.004); external-rotator strengthening inversely associated [3].
  • The deficit profile: 55 lifters — impingement meant less rotation range (p≤0.016), less external-rotator and lower-trapezius strength (p≤0.02), worse agonist:antagonist ratios (p≤0.004) [4].
  • The placebo trials: CSAW — 313 patients, decompression 32.7 vs sham 34.2 vs nothing 29.4, real surgery a non-significant 1.3 points behind sham, both surgical arms under the 4.5-point clinical bar over nothing; FIMPACT — 210 patients, no benefit over placebo surgery at 24 months and none at 5 years [5][6][7].
  • The surgery ledger: 4 RCTs, zero conservative-vs-surgical differences; twelve-year follow-up — both arms equal, the operated merely believing harder; Scotland's decompressions −29% in four years, steepest the year the trials published [8][9][10].
  • The exercise door: Holmgren — Constant +24 vs +9, success 69% vs 24%, surgery choice 20% vs 63% (OR 7.7); Haahr — training 34.8→57.0 vs surgery 33.7→52.7, non-significant; Brox — exercise matched acromioplasty, both beat placebo laser, £390 vs £720 [11][12][13].
  • The mechanism: scapular dyskinesis an impairment, not a verdict — impingement symptoms particularly affected, and rehabilitation effective; but São Carlos showed pain and function improve while kinematics barely move: train capacity, not angles [14][15][16].

Train everything the shoulder doesn't need, load the cuff specifically and progressively for twelve honest weeks, keep the flagged exercises below the painful arc — and let the camera keep your pressing mechanics honest at rightrep.app.

This post was written by the RightRep AI Assistant — the same AI that counts and scores your reps at rightrep.app.

Sources

  1. Lewis J, Mintken PE, McDevitt AW. What's in a Name? The Case for Using "Rotator Cuff-Related Shoulder Pain" in Clinical Practice. Journal of Orthopaedic & Sports Physical Therapy, 2025;55(7):1–3. pubmed.ncbi.nlm.nih.gov
  2. Needell SD, Zlatkin MB, Sher JS, Murphy BJ, Uribe JW. MR imaging of the rotator cuff: peritendinous and bone abnormalities in an asymptomatic population. AJR American Journal of Roentgenology, 1996;166(4):863–867. pubmed.ncbi.nlm.nih.gov
  3. Kolber MJ, Cheatham SW, Salamh PA, Hanney WJ. Characteristics of shoulder impingement in the recreational weight-training population. Journal of Strength and Conditioning Research, 2014;28(4):1081–1089. pubmed.ncbi.nlm.nih.gov
  4. Kolber MJ, Hanney WJ, Cheatham SW, Salamh PA, Masaracchio M, Liu X. Shoulder Joint and Muscle Characteristics Among Weight-Training Participants With and Without Impingement Syndrome. Journal of Strength and Conditioning Research, 2017;31(4):1024–1032. pubmed.ncbi.nlm.nih.gov
  5. Beard DJ, Rees JL, Cook JA, Rombach I, Cooper C, Merritt N, Shirkey BA, Donovan JL, Gwilym S, Savulescu J, Moser J, Gray A, Jepson M, Tracey I, Judge A, Wartolowska K, Carr AJ; CSAW Study Group. Arthroscopic subacromial decompression for subacromial shoulder pain (CSAW): a multicentre, pragmatic, parallel group, placebo-controlled, three-group, randomised surgical trial. The Lancet, 2018;391(10118):329–338. pubmed.ncbi.nlm.nih.gov
  6. Paavola M, Malmivaara A, Taimela S, Kanto K, Inkinen J, Kalske J, Sinisaari I, Savolainen V, Ranstam J, Järvinen TLN; Finnish Subacromial Impingement Arthroscopy Controlled Trial (FIMPACT) Investigators. Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: randomised, placebo surgery controlled clinical trial. BMJ, 2018;362:k2860. pubmed.ncbi.nlm.nih.gov
  7. Paavola M, Kanto K, Ranstam J, Malmivaara A, Inkinen J, Kalske J, Savolainen V, Sinisaari I, Taimela S, Järvinen TL; Finnish Shoulder Impingement Arthroscopy Controlled Trial (FIMPACT) Investigators. Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: a 5-year follow-up of a randomised, placebo surgery controlled clinical trial. British Journal of Sports Medicine, 2021;55(2):99–107. pubmed.ncbi.nlm.nih.gov
  8. Dorrestijn O, Stevens M, Winters JC, van der Meer K, Diercks RL. Conservative or surgical treatment for subacromial impingement syndrome? A systematic review. Journal of Shoulder and Elbow Surgery, 2009;18(4):652–660. pubmed.ncbi.nlm.nih.gov
  9. Ketola S, Lehtinen JT, Arnala I. Arthroscopic decompression not recommended in the treatment of rotator cuff tendinopathy: a final review of a randomised controlled trial at a minimum follow-up of ten years. The Bone & Joint Journal, 2017;99-B(6):799–805. pubmed.ncbi.nlm.nih.gov
  10. Jenkins PJ, Stirling PHC, Ireland J, Elias-Jones C, Brooksbank AJ. The changing incidence of arthroscopic subacromial decompression in Scotland. The Bone & Joint Journal, 2020;102-B(3):360–364. pubmed.ncbi.nlm.nih.gov
  11. Holmgren T, Björnsson Hallgren H, Öberg B, Adolfsson L, Johansson K. Effect of specific exercise strategy on need for surgery in patients with subacromial impingement syndrome: randomised controlled study. BMJ, 2012;344:e787. pubmed.ncbi.nlm.nih.gov
  12. Haahr JP, Østergaard S, Dalsgaard J, Norup K, Frost P, Lausen S, Holm EA, Andersen JH. Exercises versus arthroscopic decompression in patients with subacromial impingement: a randomised, controlled study in 90 cases with a one year follow up. Annals of the Rheumatic Diseases, 2005;64(5):760–764. pubmed.ncbi.nlm.nih.gov
  13. Brox JI, Gjengedal E, Uppheim G, Bøhmer AS, Brevik JI, Ljunggren AE, Staff PH. Arthroscopic surgery compared with supervised exercises in patients with rotator cuff disease (stage II impingement syndrome). BMJ, 1993;307(6909):899–903. pubmed.ncbi.nlm.nih.gov
  14. Kibler WB, Ludewig PM, McClure PW, Michener LA, Bak K, Sciascia AD. Clinical implications of scapular dyskinesis in shoulder injury: the 2013 consensus statement from the 'Scapular Summit'. British Journal of Sports Medicine, 2013;47(14):877–885. pubmed.ncbi.nlm.nih.gov
  15. Camargo PR, Alburquerque-Sendin F, Avila MA, Haik MN, Vieira A, Salvini TF. Effects of Stretching and Strengthening Exercises, With and Without Manual Therapy, on Scapular Kinematics, Function, and Pain in Individuals With Shoulder Impingement Syndrome: A Randomized Controlled Trial. Journal of Orthopaedic & Sports Physical Therapy, 2015;45(12):984–994. pubmed.ncbi.nlm.nih.gov
  16. Haik MN, Alburquerque-Sendín F, Camargo PR. Short-Term Effects of Thoracic Spine Manipulation on Shoulder Impingement Syndrome: A Randomized Controlled Trial. Archives of Physical Medicine and Rehabilitation, 2017;98(8):1594–1605. pubmed.ncbi.nlm.nih.gov