Orthopaedics & Traumatology: Surgery & Research (2011) 97, S195—S203

Available online at

www.sciencedirect.com

ORIGINAL ARTICLE

Pain management in the rehabilitation of stiffshoulder: Prospective multicenter comparativestudy of 193 cases

P. Gleyzea,g,∗, P.-H. Flurina, E. Laprellea, D. Katzb, B. Toussaintc,T. Benkalfated, N. Solignace, C. Lévignef, French Arthroscopy Society

a Orthopedic and Sports Surgery Center, 33700 Merignac, Franceb Clinique du Ter, 56270 Ploemeur, Francec Clinique Générale, 74000 Annecy, Franced Clinique de la Sagesse, 35000 Rennes, Francee Ambroise-Paré University Hospital Center, 92100 Billancourt, Francef Clinique du Parc, 69000 Lyon, Franceg Albert-Schweitzer Hospital, 301, avenue d’Alsace, 68000 Colmar, France

Accepted: 7 September 2011

KEYWORDSStiff shoulder;Treatment;Rehabilitation;Self-rehabilitation;Capsulotomy;Pain management;Adhesive capsulitis;Frozen shoulder;Complex regionalpain syndrome;Therapeuticeducation

SummaryIntroduction: The present study investigated the impact of respecting pain threshold on clinicalrecovery in stiff shoulder.Patients and methods: A prospective multicenter comparative study followed up 193 casesof shoulder stiffness for a mean 12-month period (range, 8—31 months) after four differ-ent treatment protocols: (1) conventional sub-pain-threshold rehabilitation (58 cases); (2)self-rehabilitation exceeding the pain threshold (59 cases); (3) supervised suprathreshold reha-bilitation (31 cases); and (4) capsulotomy with sub-threshold rehabilitation (45 cases). Follow-upwas daily for the first 6 weeks then weekly for the next 6; each session included assessment ofthe painfulness, feasibility and duration of each rehabilitation and self-rehabilitation exerciseand of pain status, disability and psychological status. The surgeon followed patients up at6 weeks, 3 months, 6 months, 1 year and at last follow-up.Results: Sub-threshold rehabilitation provided progressive results, limited in time (P < 0.05).Suprathreshold self-rehabilitation provided reduced pain (P < 0.05) as of the first days, with noc-

ays’ rehabilitation in 43% of cases. Supervision of self-rehabilitation

turnal pain ceasing after 7 d exercises optimized the clinical result (P < 0.05). Capsulotomy did not influence pain evolutionover the first 8 weeks, but then improved it. Failure (at 1 year, 14—17%; last follow-up, 3.5%)correlated directly with the number of exercises performed by the patient (P < 0.05).

∗ Corresponding author. Tel: +(33) 3 89 23 09 90; fax: +(33) 3 89 29 05 94.E-mail address: pascal.gleyze@orange.fr (P. Gleyze).

1877-0568/$ – see front matter © 2011 Elsevier Masson SAS. All rights reserved.doi:10.1016/j.otsr.2011.09.006

S196 P. Gleyze et al.

Discussion: The dogma of respecting the pain threshold is dated: pain inflicted on a passivepatient impairs clinical evolution, but pain managed by an informed active patient underexperienced supervision provides rapid recovery of function and pain-free status.Level of evidence: Level III, case-control, prospective comparative.© 2011 Elsevier Masson SAS. All rights reserved.

I

Petcbonrifctvcap

uos

rtciocpb

emr

P

A1tw

ordstr

d

p6tm(

cfratt[bpn

Cn

Pr(wbc2

S

Psettobt

Supervised suprathreshold self-rehabilitation(SSSR: n = 31)

ntroduction

ain is an extra item on our care list; it is a fact, experi-nced by the patient but often difficult for the physiciano deal with, and which strongly impacts every aspect ofare, partly determining its conditions and results. Pain felty the patient may affect our decisions and the quality ofur treatment processes and, thus, of our results. Pain doesot exist in itself: it is the painful experience that is theeality we have to deal with. ‘‘Pain is something else, it isnhuman’’ [1,2]. Pain is the felt result of a state of sufferingor which we can only look for the mental or physiologi-al conditions that have given rise to it [3]. The problem iso integrate this subjective datum, with its great potentialariability, into the objective and statistical management oflinical assessment [4]. Rehabilitation exercises are painful,nd our everyday practice works with the paradox of causingain in order to alleviate it.

History in general, and the history of medicine in partic-lar, shows that any unknown generates fear, and methodsf care based on doubt and fear of complications may proveelf-defeating [4—7].

A perfect example of this is the dogma, in stiff shoulderehabilitation, of respecting the pain threshold, founded onhe principle of ‘‘Above all, do no harm!’’ [8], out of fear ofomplications believed to correlate with pain [9—19]. Thiss all the more important as the onset mechanisms, etiol-gy and nosology of complications (capsulitis, stiff shoulder,omplex regional pain syndrome, frozen shoulder, etc.) areoorly known and their management remains empirical andeset by lack of knowledge [20—26].

The present study, therefore, examined the impact onxperienced pain and on clinical outcome of different treat-ent options in stiff shoulder according to whether they

espect the pain threshold or not.

atients and method

prospective multicenter study was performed on a series of93 cases of stiff shoulder, comparing subjective and objec-ive clinical evolution per treatment option according tohether the pain threshold is respected or not.

All patients presented with significantly reduced rangef passive motion (passive forward flexion < 150◦, externalotation more than 20◦ less than in the contralateral shoul-er and reduced internal rotation). History of surgery fortiff shoulder, degenerative bone lesions and fractures less

han 3 months old or non-consolidated were exclusion crite-ia.

Clinical and pain assessment included normalized shoul-er examination in dorsal decubitus [27—29] and was

Fwia

erformed on the day of inclusion and at 6 weeks, 3 months,months, 1 year and at last follow-up, with analysis of his-

ory and of onset and objective (goniometry of ranges ofotion) and subjective (Constant score) assessment [30]

Appendix 1: follow-up form).Each rehabilitation session was assessed by either a physi-

ian or a physiotherapist: feasibility, pain level and durationor each conventional rehabilitation exercise (Appendix 2:ehabilitation follow-up form). The patients also filled outsubjective questionnaire, daily for the first 6 weeks and

hen weekly up to the 3rd month, scoring daytime and night-ime pain, functional disturbance and morale (for pain: 0no pain] to 10 [maximal pain]; disturbance: 0 = no distur-ance; morale: 0 = morale at its lowest). At the same times,atients undergoing self-rehabilitation filled out a question-aire assessing each exercise on the same criteria.

Four treatment groups were distinguished.

onventional sub-threshold rehabilitation (CSR:= 58)

atients were managed by conventional rehabilitation,especting a pain threshold defined by a visual analog scaleVAS) score < 6/10. They had three to five sessions a weekith a physiotherapist, in some cases working under a reha-ilitation physician. The usual methods and exercises ofonventional rehabilitation were applied [9,27] (Appendix: rehabilitation follow-up form).

uprathreshold self-rehabilitation (SSR: n = 59)

atients were managed by self-rehabilitation (n = 59) orupervised self-rehabilitation (SSSR, below: n = 31) andncouraged to go over their pain threshold (i.e., over six onhe VAS). They were asked to be reasonable and to dividehe daily exercise load into 5—10-minute sessions spreadver the day [27,28,31]. The exercises were simple ones,ased on everyday movements [31]. The patients were lefto themselves between the follow-up consultations.

or this group, self-rehabilitation was backed up by sessionsith a physiotherapist trained to include the idea of exceed-

ng the pain threshold in the communication with the patientnd in the physiotherapy treatment [27].

pttccrof(

Cr

TtbefWoSl(

afiwt

I

TtbS

S

Ial(digC

plirdgs

F

Pain management in the rehabilitation of stiff shoulder

Endoscopic circumferential capsulotomy withmobilization under anesthesia and conventionalsub-threshold rehabilitation (CAPS: n = 45)

The capsulotomy technique was classical [32—34] and reha-bilitation was as in the CSR group.

An on-line database was set up (Carl BiostatisticTM) withon-site assessment form data entry.

A comparative descriptive study was performed pertreatment group and per technique (analysis of variance,t and Chi2 tests; significance threshold, P < 0.05) followedby a study of correlations between treatments and betweenassessment parameters (simple regression or bivariate cor-relations; significance threshold, P < 0.05). Pain levels werestudied on factorial analysis.

Results

One hundred and ninety-three patients were followed upfor a mean 13 months (range, 3—21 months) after a mean of12 months’ evolution (range, 8—31 months). Mean age was50 years (range, 18—71 years) and the mean interval to inclu-sion 15 months (range, 5—28 months). The sex ratio was 62%female. Fifty percent of cases were considered spontaneousstiffness, with a mean interval to inclusion of 12 months, ver-sus 3 months for post-traumatic stiffness (31%) and 9 monthsfor stiffness secondary to surgery (12%).

There were no significant differences in baseline painintensity between treatment groups. Pain intensity corre-lated directly with degree of stiffness in elbow-to-bodyexternal rotation [29] and forward flexion (P < 0.05).

Comparative descriptive functional evolution

Fig. 1 compares evolution of overall function (Constantscore) between treatment groups. Functional recovery onConstant score systematically correlated with mean painintensity (P < 0.05) and degree of stiffness (P < 0.05). On fac-torial analysis, the pain criterion accounted for 83% of theoverall Constant score (out of 100 points), although only 15of the 100 points directly assess pain. This demonstrates theintercorrelation between pain and the objective and subjec-tive factors of the Constant score as a whole.

Conventional sub-threshold rehabilitation

The CSR group presented with moderate to severe baselinepain (graded 6/10); day- and night-time pain, then progres-sively diminished but rose again significantly (P > 0.05) afterthe 12th week, which usually corresponded to the time whenthe physiotherapist began to reduce treatment. Day- andnight-time pain showed parallel evolution (Fig. 2). Fig. 3shows the relation between functional improvement andalleviation of pain, with persistent moderate pain limitingfunctional recovery (P < 0.05) over periods beyond 6 months.

Suprathreshold self-rehabilitation

In the SSR group, the evolution of day- versus night-time paindiffered, with a large rapid decrease in the latter: 43% of

Fea

S197

atients free of night-time pain after 7 days (P < 0.05). Day-ime pain diminished strongly during the first days (P < 0.05)hen stabilized, with a background of persistent pain directlyorrelating to exercise duration (Fig. 4). Pain alleviationan be seen to stagnate between weeks 6 and 12, cor-esponding to a reduction in exercises and a lower ratef functional improvement (P < 0.05), after which shoulderunction improved in parallel to alleviation of pain (P < 0.05)Fig. 5).

omparison between pain evolution according toespect of the pain threshold entre (CSR vs SSR)

he SSR group showed less pain than the CSR group duringhe first days (P < 0.05). Thereafter, the exercises performedy the patient may sustain a background of daytime painqual to or greater than in the CSR group, but with greaterunctional improvement during the first 6 weeks (P < 0.05).hen the frequency of exercise dropped, around week 12

f self-rehabilitation, the background of daytime pain in theSR group diminished and mean pain intensity fell below theevel of the CSR group and remained so until last follow-upP < 0.05) (Fig. 6).

The SSSR group, supervised by a physiotherapist, showedpattern of evolution similar to the SSR group’s for the

rst 6 weeks, but with greater alleviation of pain betweeneeks 6 and 12 (P < 0.05), and becoming again equivalent to

he SSR level thereafter.

nfluence of capsulotomy on pain

he evolution of pain in the CAPS (capsulotomy) group washe same as in the CSR group for the first 8 weeks, thenecoming significantly better (P < 0.05), at the level of theSR and SSSR groups (Fig. 7).

ynthesis of correlations

n the groups treated with respect of the pain threshold (CSRnd CAPS), daytime pain levels correlated with night-timeevels (P < 0.05), disturbance (P < 0.05) and impaired moraleP < 0.05). They were proportional to rehabilitation sessionuration and showed a systematic negative impact, decreas-ng after week 6, depending on the exercises performed, ononiometric range of motion and subjective and objectiveonstant criteria (P < 0.05), and thus on the clinical result.

In the groups treated by self-rehabilitation exceeding theain threshold (SSR and SSSR), daytime pain levels corre-ated with exercise intensity and duration (P < 0.05). Unliken the CSR and CAPS groups, daytime pain intensity cor-elated significantly with improvement in night-time pain,isturbance and morale (P < 0.05), with positive impact ononiometric range of motion and function on the Constantcore (P < 0.05) at whatever time of follow-up.

ailure

ailure criteria were defined as anterior elevation < 140◦,xternal rotation more than 20◦ less than contralaterallynd Constant score < 70 points. Failure rates were identical

S198 P. Gleyze et al.

Figure 1 Functional evolution according to treatment group. Constant score/100, week of FU.

Figure 2 Evolution of daytime and night-time pain under conventional sub-threshold rehabilitation.

Figure 3 Evolution of function and pain under conventional sub-threshold rehabilitation. Function/100% in blue (100%: maximalfunction), pain/100% in red (100%: maximal pain).

Figure 4 Dissociation of day- and night-time pain under suprathreshold self-rehabilitation. (10: maximal pain).

Figure 5 Evolution of function and pain under suprathreshold self-rehabilitation. Function/100% in blue (100%: maximal function),pain/100% in red (100%: maximal pain).

Figure 6 Evolution of mean pain under conventional sub-threshold rehabilitation and suprathreshold self-rehabilitation. (Maximalpain: 10).

Pain management in the rehabilitation of stiff shoulder S199

n the

at

ptlswadwnate

tapto

caTsmt

fetctfctt

C

Suatft

Figure 7 Evolution of mean pain i

with sub- or suprathreshold rehabilitation at 1 year (CSR,14%; SSR, 17%) and at end of follow-up (CSR, 0%; SSR, 7%).Overall functional score (Constant score < 70) was lower withsub-threshold rehabilitation at 1 year (P < 0.05) (CSR, 26%failure; SSR, 5.9%) but showed no significant difference atend of follow-up. The failure rate under sub-threshold reha-bilitation correlated at 1 year with the level of the patient’spersonal work and with pain intensity at final follow-up(P < 0.05). The capsulotomy group showed no failures forwhatever criterion, time point or pain level (P < 0.05).

Discussion

There have been few studies of pain management inshoulder pathology [9,14,31,34] and follow-up durations infollow-up studies made it difficult to assess the specific clin-ical impact of treatment modalities [10,35—39].

The multicenter design of the present study may havebiased the collection of certain results (subjective Constantscores), but the simple follow-up criteria (pain, disturbance,morale) were easy for the patient to assess, making thefollow-up data reliable, sensitive and statistically inter-pretable.

Conventional rehabilitation in which patient and ther-apist take care to respect the pain threshold provedmore painful at the beginning of treatment than self-rehabilitation with encouragement to exceed the threshold.During the following 6 weeks, conventionally managedpatients suffered a little less than self-rehabilitatingpatients, but that was due to the intensity of the exercisesperformed by the latter. The therapeutic power of personalwork has been proved [40—42] and exercise intensity andits role in recovering muscle force is known to reduce painand risk of dysfunction due to subacromial impingement[14,43—45].

The resurgence of pain when conventional rehabilita-tion tails off or ceases after a few months is a well-knownfact [24,46,47], which we see as symptomatic of therapistdependence and lack of therapeutic education to sustainthe results achieved with the physiotherapist. The cultureof fear of pain and of the nebula of possible complications,usually unclearly understood by both patient and therapist[9,13,14,48], seem to contribute to this end-of-treatmentresurgence of pain.

Suprathreshold self-rehabilitation provides rapid relief,especially of night-time resurgence of pain. Residual painis mainly daytime and directly correlated with the duration

of the exercises performed by the patient, who is usuallywell motivated, reassured and exercising in a panic-freecontext [1—49]. The instructions at the very least restore amore peaceful daily life with control over stiffness-inducing

pat

three groups. (Maximal pain: 10).

nti-pain reflexes and particularly over reflex trapezoid con-ractures that induce or prolong shoulder stiffness [9,28,31].

The self-rehabilitation exercise form is a material sup-ort, which may be considered symbolic in terms of medicalechnicality but the main role of which is to give the patienticense to treat him or herself [1,31]. The present studyhows that patients who work on and manage their exercisesithout fear of pain or complications achieve an immedi-tely improved quality of life in terms of pain and morale, inirect proportion to their degree of personal commitment,ith functional results that are both rapid and sustainable,ot being dependent on therapists. The paradox is thus toccept and get the patient to accept a ‘‘liberating’’ painhat is efficacious only when associated to clear therapeuticducation.

Supervision of suprathreshold self-rehabilitation by arained physiotherapist, able to communicate soberlybout complications and willing to involve the patient inain-inducing but supervised personal exercises to whichreatment will be adapted, avoids the discouraging effectf being too alone in managing one’s treatment [40,47,48].

Capsulotomy provides no extra benefit in terms of painompared to self-rehabilitation [50], but ensures pain reliefnd non-recurrence as of the 3rd month of treatment.he heavy successive treatments entailed by capsulotomyeverely medicalize care, but with long-term benefit thatay make it indicated, in certain cases resistant, to other

reatments.Failure analysis directly implicated patient commitment:

ailure was associated with insufficient or defective personalxercise, showing the major role of active involvement andhus of therapeutic education in pain relief and clinical out-ome. Less active patients suffer greater pain and should,herefore, be considered partially responsible for their ownailure. Reassuring management of the pain threshold avoidsomplications such as complex regional pain syndrome, andhe failure rate is similar to that in conventional rehabilita-ion.

onclusion

houlder stiffness is a pathology that needs to be followedp for more than 1 year and in which experienced pain issignificant prognostic factor. The present study showed

hat patients undergoing sub-threshold rehabilitation suf-er greater pain, with poorer results and lower morale thanhose performing self-rehabilitation.

We, therefore, consider that the tolerance threshold forain inflicted on a passive and frightened patient is low,nd impairs clinical evolution. In contrast, the tolerancehreshold for an active and reassured patient is higher, which

S

eit

iaw‘tp

wb

D

T

200

nhances efficacy in performing exercises and improves clin-cal evolution in as much as the patient is managing a painhat has been explained and rendered less frightening.

The dogma of respecting the threshold for painnflicted on the patient should give way to the idea of

‘‘reasonable’’ pain threshold managed by a patientho is aware and involved. We need to move on from

‘respecting the pain threshold’’ to ‘‘managing the painhreshold’’, holding to the essential principles of thera-eutic education and active patient involvement, without

c

A

P. Gleyze et al.

hich therapeutic and analgesic success is not possi-le.

isclosure of interest

he authors declare that they have no conflicts of interestoncerning this article.

ppendix 1. Follow-up form

S201

Pain management in the rehabilitation of stiff shoulder

Appendix 1(Continued )

S P. Gleyze et al.

A

R

202

ppendix 2. Rehabilitation follow-up form

eferences

[1] Canguilhem G. Le normal et le pathologique, Quadrige. 3e ed.Presses universitaires de France: 1993.

[2] Debru C, Canguilhem G. Science et non science. Éditions Rued’Ulm / Presses de l’École normale supérieure; 2004.

[3] Lalande A. Vocabulaire technique et critique de la philosophie.

17e ed. Presses universitaires de France: 1991.

[4] Gleyze P. La donnée en recherche clinique, de la philosophiedes sciences à l’analyse statistique. Bull Hist Epistem Sc Vie1999;6(1):47—54.

[5] Sun Tzu. L’art de la guerre (The Art of War). Paris: HachetteLittératures; 2000.

[6] Changeux JP. Du vrai, du beau, du bien. Une nouvelle approcheneuronale. Odile Jacob: 2008.

[7] Monod J. Le hasard et la nécessité, Essai sur la philosophienaturelle de la biologie moderne. Éditions du Seuil: 1970.

[8] Smith CM. Origin and uses of primum non nocere — above all,do no harm! J Clin Pharmacol 2005;45(4):371—7.

[9] Ellenbecker TS, Cools A. Rehabilitation of shoulder impinge-ment syndrome and rotator cuff injuries: an evidence-basedreview. Br J Sports Med 2010;44(5):319—27.

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

Pain management in the rehabilitation of stiff shoulder

[10] Binder AI, Bulgen DY, Hazleman BL, Roberts S. Frozenshoulder: a long-term prospective study. Ann Rheum Dis1984;43(3):361—4.

[11] Codman EA. The shoulder: rupture of the supraspinatus tendonand other lesions in or about the subacromial bursa. Boston,MA: T Todd Company; 1934.

[12] Cyriax JH. Text-book of Orthopædic Medicine. London: Cassell;1962.

[13] Diercks RL, Stevens M. Gentle thawing of the frozen shoulder:a prospective study of supervised neglect versus inten-sive physical therapy in 77 patients with frozen shouldersyndrome followed up for 2 years. J Shoulder Elbow Surg2004;13(5):499—502.

[14] Green S, Buchbinder R, Hetrick S. Physiotherapy inter-ventions for shoulder pain. Cochrane Database Syst Rev2003;(2):CD004258.

[15] Grey RG. The natural history of ‘‘idiopathic’’ frozen shoulder.J Bone Joint Surg Am 1978;60(4):564.

[16] Neviaser RJ, Neviaser TJ. The frozen shoulder: diagnosis andmanagement. Clin Orthop Relat Res 1987;223:59—64.

[17] Lloyd-Roberts GC, French PR. Periarthritis of the shoul-der: a study of the disease and its treatment. Br Med J1959;1(5137):1569—71.

[18] Patte, Goutallier D. De quelques réflexions sur le traitementchirurgical des lésions de la coiffe des rotateurs. Cahiersd’enseignement de la SOFCOT, L’épaule douloureuse chirurgi-cale 1988;33:113—22.

[19] Duplay S. De la périarthrite scapulohumérale et des raideursde l’épaule qui en sont la conséquence. Arch Gen Med1872;513:542.

[20] Neviaser AS, Hannafin JA. Adhesive capsulitis: a review of cur-rent treatment. Am J Sports Med 2010;38:2346.

[21] Neviaser JS. Adhesive capsulitis of the shoulder. J Bone JointSurg Am 1945;27(2):211—22.

[22] Reeves B. The natural history of the frozen shoulder syndrome.Scand J Rheumatol 1975;4(4):193—6.

[23] Rill BK, Fleckenstein CM, Levy MS, Nagesh V, Hasan SS. Predic-tors of outcome after non operative and operative treatmentof adhesive capsulitis. Am J Sports Med 2011;39(3):567—74.

[24] Griggs SM, Ahn A, Green A. Idiopathic adhesive capsulitis: aprospective functional outcome study of non operative treat-ment. J Bone Joint Surg Am 2000;82(10):1398—407.

[25] Hand C, Clipsham K, Rees JL, Carr AJ. Long-term outcome offrozen shoulder. J Shoulder Elbow Surg 2008;17(2):231—6.

[26] Hsu JE, Anakwenze OA, Warrender WJ, Abboud JA. Cur-rent review of adhesive capsulitis. J Shoulder Elbow Surg2011;20(3):502—14.

[27] Flurin PH, Laprelle E, Bentz JY, Lachaud C, Boy M,Pellet JL, et al. Rééducation de l’épaule non opérée.Encyclopédie médicochirurgicale, kinésithérapie-médecinephysique-réadaptation [26-210-B-10].

[28] Coudane H, Gleyze P, Thierry G, Charvet R, Blum A. Raideurde l’épaule. Encyclopédie médicochirurgicale. Appareil loco-moteur [14-352-A-10].

[29] Wolf EM, Cox WK. The external rotation test in the diagnosisof adhesive capsulitis. Orthopedics 2010;33(5):303.

[30] Constant C, Murley AH. Clinical method of functional assess-ment of the shoulder. Clin Orthop 1987;214:160—4.

[31] Gleyze P, Tamisier A, Bauer P. Protocole d’autorééducation for-cée de l’épaule : présentation de la fiche pratique et résultats

préliminaires. Rev Chir Orthop Traumatol 2003;89(2):167—82.

[32] Ide J, Takagi K. Early and long-term results of arthro-scopic treatment for shoulder stiffness. J Shoulder Elbow Surg2004;13(2):174—9.

[

S203

33] Jerosch J. 360◦ arthroscopic capsular release in patients withadhesive capsulitis of the glenohumeral joint: indication, sur-gical technique, results. Knee Surg Sports Traumatol Arthrosc2001;9(3):178—86.

34] Roosink M, Renzenbrink GJ, Buitenweg JR, Van Dongen RT,Geurts AC, Ijzerman MJ. Persistent shoulder pain in the first6 months after stroke: results of a prospective cohort study.Arch Phys Med Rehabil 2011;92(7):1139—45.

35] Cheville AL, Tchou J. Barriers to rehabilitation follow-ing surgery for primary breast cancer. J Surg Oncol2007;95(5):409—18.

36] Andersen NH, Sojbjerg JO, Johannsen HV, Sneppen O. Frozenshoulder: arthroscopy and manipulation under general anes-thesia and early passive motion. J Shoulder Elbow Surg1998;7(3):218—22.

37] Beaufils P, Prévot N, Boyer T, et al. Arthroscopic release of theglenohumeral joint in shoulder stiffness: a review of 26 cases.French Society for Arthroscopy. Arthroscopy 1999;15:49—55.

38] Bulgen DY, Binder AI, Hazleman BL, Dutton J, Roberts S. Frozenshoulder: prospective clinical study with an evaluation of threetreatment regimens. Ann Rheum Dis 1984;43(3):353—60.

39] Miller MD, Wirth MA, Rockwood Jr CA. Thawing thefrozen shoulder: the ‘‘patient’’ patient. Orthopedics1996;19(10):849—53.

40] Kivimaki J, Pohjolainen T, Malmivaara A, et al. Manipulationunder anesthesia with home exercises versus home exercisesalone in the treatment of frozen shoulder: a randomized,controlled trial with 125 patients. J Shoulder Elbow Surg2007;16(6):722—6.

41] Light KE, Nuzik S, Personius W, Barstrom A. Low-load prolongedstretch vs high-load brief stretch in treating knee contractures.Phys Ther 1984;64(3):330—3.

42] Tanaka K, Saura R, Takahashi N, Hiura Y, Hashimoto R. Jointmobilization versus self-exercises for limited glenohumeraljoint mobility: randomized controlled study of management ofrehabilitation. Clin Rheumatol 2010;29(12):1439—44.

43] Celik D, Sirmen B, Demirhan M. The relationship of musclestrength and pain in subacromial impingement syndrome. ActaOrthop Traumatol Turcica 2011;45(2):79—84.

44] Ellenbecker TS, Davies GJ, Rowinski MJ. Concentric ver-sus eccentric isokinetic strengthening of the rotator cuff.Objective data versus functional test. Am J Sports Med1998;16(1):64—9.

45] Ludewig PM, Cook TM. Alterations in shoulder kinematics andassociated muscle activity in people with symptoms of shoulderimpingement. Phys Ther 2000;80(3):276—91.

46] Vermeulen HM, Obermann WR, Burger BJ, Kok GJ, Rozing PM,van Den Ende CH. End-range mobilization techniques in adhe-sive capsulitis of the shoulder joint: a multiple-subject casereport. Phys Ther 2000;80(12):1204—13.

47] Vermeulen HM, Rozing PM, Obermann WR, Le Cessie S,Vliet Vlieland TP. Comparison of high-grade and low-grademobilization techniques in the management of adhesive cap-sulitis of the shoulder: randomized controlled trial. Phys Ther2006;86(3):355—68.

48] Quraishi NA, Johnston P, Bayer J, Crowe M, Chakrabarti AJ.Thawing the frozen shoulder: a randomised trial comparingmanipulation under anaesthesia with hydrodilatation. J BoneJoint Surg Br 2007;89(9):1197—200.

49] Allamel-Raffin C, Gangloff JL. La raison et le réel. Champs

philosophiques. Paris: Ellipes Éditions; 2007.

50] Nicholson GP. Arthroscopic capsular release for stiff shoul-ders: effect of etiology on outcomes. Arthroscopy 2003;19(1):40—9.