Mobilization Techniques With Frozen Shoulder

Mobilization Techniques for Frozen Shoulder SyndromeSubjectsSubjects with FSS were recruitedfrom the clinics in the Department ofPhysical Medicine and Rehabilitationat National Taiwan University Hospital. Based on the judgment of whatconstitutes clinically meaningful differences and variability estimatesfrom previous studies,3,12,21,22 a sample size of 15 subjects per group provided 80% power to detect differences of 5 degrees of ROM betweenthe preintervention and postintervention measurements as well as between the 2 groups of interest at analpha level of .05 with a 2-tailed test.The sample size estimate shouldbe based on functional outcome asa standard to assess the effect ofintervention. Variability, lack of reliability, or not enough sensitivityof functional outcome assessmentsin previous studies, however, precluded our use of a functional statusmeasure. Thus, we used ROM to determine the sample size in our study.The participants received writtenand verbal explanations of the purposes and procedures of the study. Ifthey agreed to participate, theysigned informed consent forms approved by the Human Subjects Committee of National Taiwan UniversityHospital.All subjects with FSS fulfilled the following inclusion criteria:(1) having a painful stiff shoulder forat least 3 months, (2) having limitedROM of a shoulder joint (ROM lossesof 25% or greater compared with thenoninvolved shoulder in at least 2 ofthe following shoulder motions: glenohumeral flexion, abduction, ormedial and lateral rotation), and(3) the consent of the subject’s physician to participate in the study. Theexclusion criteria were: (1) diabetesmellitus, (2) a history of surgery onthe particular shoulder, (3) rheumatoid arthritis, (4) a painful stiffshoulder after a severe trauma,(5) fracture of the shoulder complex,(6) rotator cuff rupture, or (7) tendon calcification.October 2007InterventionsParticipants in both groups receivedmobilization treatments twice aweek for 30 minutes and a simpleexercise program comprising pendular exercises and scapular setting(isometric scapular retraction). Aphysical therapist with 8 years ofclinical experience in manual therapy provided the intervention. Noother interventions—including physical modalities (ie, ultrasound, shortwave diathermy, and electrotherapy), intra-articular steroid injection,or arthrographic joint distension—were allowed for the duration of thetrial. The subjects were not instructed in home exercises in orderto exclude the influence of their adherence to the exercise protocol.Additionally, frequent remindersduring instruction and telephonecalls were given to the subjects topersuade them not to do homeexercises.Mid-Range MobilizationAn MRM technique was performedon the involved shoulder, as described by Maitland17 and Kaltenborn.18 With the subject in a relaxedsupine position, the humerus wasmoved to the resting position (40 ofabduction). While the humerus washeld in this position, 10 to 15 repetitions of the mobilization techniques were applied.End-Range MobilizationIn addition to the MRM technique,ERM has been recommended.3,16,17The intent of ERM was not only torestore joint play but also to stretchcontracted periarticular structures.We used the techniques describedby Vermeulen et al3 and Maitland17as follows. At the start of each intervention session, the physical therapist examined the subject’s ROM toobtain information about the endrange position and the end-feel ofthe glenohumeral joint. Then, thetherapist’s hands were placed closeto the glenohumeral joint, and theVolume 87humerus was brought into a positionof maximal range in different directions. Ten to 15 repetitions of intensive mobilization techniques, varying the plane of elevation or varyingthe degree of rotation in the endrange position, were applied.Mobilization With MovementThe use of MWM for peripheral jointswas developed by Mulligan.19,20 Thistechnique combines a sustained application of a manual technique “gliding”force to a joint with concurrent physiologic (osteo-kinematic) motion ofthe joint, either actively performed bythe subject or passively performedby the therapist. The manual force, ormobilization, is theoretically intendedto cause repositioning of bone positional faults. The intent of MWM is torestore pain-free motion at joints thathave painful limitation of range ofmovement.The MWM technique was performedon the involved shoulder as described by Mulligan.19,20 With thesubject in a relaxed sitting position, abelt was placed around the head ofthe humerus to glide the humerushead appropriately, as the therapist’shand was used over the appropriateaspect of the head of the humerus. Acounter pressure also was applied tothe scapula with the therapist’sother hand. The glide was sustainedduring slow active shoulder movements to the end of the pain-freerange and released after return to thestarting position. Three sets of 10repetitions were applied, with 1minute between sets.Outcome AssessmentDisability assessment. The Flexilevel Scale of Shoulder Function(FLEX-SF) is a self-administered,shoulder-specific, fixed-item indexconsisting of 3 levels of function. Inthis scale, respondents answer a single item that grossly classifies theirlevel of function as low, medium, orhigh.26 They then respond only toNumber 10Physical Therapy f1309

Mobilization Techniques for Frozen Shoulder Syndromethe items that targeted their level offunction. Scores are recorded from1, indicating the most limited function, to 50, indicating the absence oflimited function in the subject. Thisscale has been shown to have highreliability (intraclass correlation coefficient [ICC] .90) and validity (responsiveness index 1.2).Shoulder complex kinematics.The FASTRAK motion analysis system* was used to record shouldercomplex kinematics. The details ofthe method can be found in our previous reports.27,28 In general, 3 sensors for the system were attached tothe bony landmarks. One sensor wasattached to the sternum, and onesensor was attached to the flat bonysurface of the scapular acromionwith adhesive tape. The third sensorwas attached to the distal humeruswith Velcro straps.†The local coordinate system developed from the digitized anatomicallandmarks for the trunk and humerus was used to describe clinicallyrelevant motions of the shoulder.Scapular orientation relative to thethorax was described using a Eulerangle sequence of rotation about Zs(protraction/retraction),rotationabout Y s (downward/upward rotation), and rotation about X s (posterior/anterior tipping). Humeral orientation relative to the thorax wasdescribed using a Euler angle sequence in which the first rotationrepresented the plane of elevation,the second rotation defined theamount of elevation, and the thirdrotation described the amount of axial rotation.Recordings started with the subjectsin a sitting position with arms relaxed at the sides. Kinematic data* Polhemus Inc, 1 Hercules Dr, PO Box 560,Colchester, VT 05446.† Velcro USA Inc, 406 Brown Ave, Manchester, NH 03103.1310fPhysical TherapyVolume 87were collected for 5 seconds in thisresting seated posture. Subjects thenwere asked to perform full activeROM in 3 tests: abduction in thescapular plane, hand-to-neck, andhand-to-scapula. Hand-to-neck andhand-to-scapula tests representedfunction-related tests.29 To determine the abduction in the scapularplane, subjects were guided to remain in the scapular plane oriented40 degrees anterior to the coronalplane. Three replicated movementswere performed in each test to themaximum possible active motions ofthe arms. The order of tests was randomized. To quantitatively characterize shoulder and scapular kinematics, the peak humeral elevationangle, the scapulohumeral rhythm(slope of scapular upward rotation toglenohumeral elevation), and thepeak scapular tilt were used as dependent variables in the abductionin the scapular-plane test. For thehand-to-neck and hand-to-scapulatests, the peak external rotationROM and peak internal rotation ROMwere used as dependent variables.All of the dependent variables werecalculated from the mean of 3 trials.Good reliability (ICC .91–.99) ofthis method has been demonstrated.28Data AnalysisAll analyses were conducted withSPSS for Windows, version 11.0.‡ Totest whether a difference of treatment efficacy existed among mobilization techniques in subjects withFSS, for each group, an analysis ofcovariance (ANCOVA) was performed using the follow-up data at 3,6, 9, and 12 weeks for each of theoutcomes, with adjustment for thebaseline values of the outcome ofinterest. To test the efficacy of 2treatments (ERM versus MWM), independent t tests were conducted tocompare change of outcome variables between 2 groups (A-B in one‡SPSS Inc, 233 S Wacker Dr, Chicago, IL60606.Number 10group versus A-C in the other groupat 6 weeks, A-C in one group versusA-B in the other group at 12 weeks).For the analysis, dropout data wereexcluded. Additionally, intention-totreat analysis was performed by including the dropout data (carryingthe last data point forward into analysis). A secondary analysis exploringthe effect of subjects dropping outwas performed using chi-square testsand survival analysis.We evaluated the potential errorswhich might affect the accuracy ofthe data. First, anthropometric variables were considered as possiblecovariates using ANCOVA, includingbody weight and body height. Second, validating sensor placementswith sensors fixed to pins embeddedin the bone, Karduna et al30 indicated that data collected from theacromion method were acceptablewhen humeral elevation stayed below 120 degrees. We compared thescapular kinematic variables by dividing the subjects into 2 groups:those with humeral elevation lessthan 120 degrees during the tasksand those with humeral elevationgreater than 120 degrees during thetasks. Third, Karduna et al30 alsofound scapular motion to be overrepresented by an average of 6 degrees when using acromion-basedsurface sensor techniques. We adjusted the data based on the assumedbias by adding 6 degrees to the humeral elevations that were greaterthan 120 degrees, which adjusted forthis error.ResultsThirty subjects were recruited andrandomly assigned to 2 groups(Tab. 1). Two subjects failed to attend the treatment. In addition, 3subjects in the A-B-A-C group werelost to follow-up because there wasno improvement during MRM treatment at 9 weeks. In the A-C-A-Bgroup, 2 subjects were lost tofollow-up because there was no imOctober 2007

Mobilization Techniques for Frozen Shoulder Syndromeprovement during MRM treatmentsat 3 weeks and 9 weeks (Fig. 1). Nosubject reported performing homeexercises.Similar results were found betweenexclusion of dropout data andintention-to-treat analysis (inclusionof dropout data). There were significant improvements (P .01) inFLEX-SF, arm elevation, scapulohumeral rhythm, humeral externalrotation, and humeral internal rotation for ERM and MWM for bothgroups. No significant improvementin outcomes was shown with MRMfor either group (Tab. 2). There wasno significant difference in outcome improvement between ERMand MWM except in scapulohumeralrhythm (Tab. 3). Mid-range mobilization corrected scapulohumeralrhythm significantly better (from0.92 to 0.68) than ERM did (from0.83 to 0.78) in subjects with FSS(Fig. 2).There were no significant differences in numbers of subjects dropping out in each group (Pearson 2 .094, P .76). A further secondary analysis was performed using survival analysis. A life table was produced using time to drop out as thesurvival variable, and comparisonswere made between the 2 groupsusing the Wilcoxon (Gehan) statistic. There also were no significantdifferences in the survival experiences of the 2 groups (value 0.035,P .851).Regarding the accuracy of the data,neither of the 2 covariates (bodyweight and body height) significantly influenced the results of theanalysis (P .05). There was no difference in the scapular kinematicvariables between the 2 groups withhumeral elevations less than orgreater than 120 degrees during thetasks (P .05). Even with the addition of the adjusted bias, neither theANCOVA nor the t-test resultsOctober 2007Table 1.Basic Characteristics of Subjects With Frozen Shoulder in the 2 Intervention Groups(n 28)aCharacteristicA-B-A-CGroup(nⴝ14)Age (y), X SD53.3 6.5Duration of symptoms (wk), X SDFemaleDominant handcFLEX-SF, X SD18 8A-C-A-BGroup(nⴝ14)58 10.1.3822 10.5613118726.8 4.4Pb28 3.7.23Arm elevation ( ), X SD106 26116 15.34Scapular tipping ( ), X SD12.7 7.910.9 7.0.160.9 0.30.8 0.3.43Humeral lateral rotation ( ), X SD45.8 16.238.2 13.6.13Humeral medial rotation ( ), X SD13.4 7.613.1 9.7.64Scapulohumeral rhythm, X SDaA mid-range mobilization, B end-range mobilization, C mobilization with movement, FLEXSF Flexilevel Scale of Shoulder Function.bDifferences in subject characteristics between the 2 groups at baseline, independent t test.cInvolved hand was dominant hand in these subjects.changed. Therefore, the placementerror is likely to have had little effecton our results.Discussion and ConclusionsOur study showed positive findings.There was an improvement in mobility and functional ability at 12 weeksin subjects treated with the 3 mobilization techniques. Comparing theeffectiveness of the 3 treatment strategies in subjects with unilateral FSS,ERM and MWM were more effectivethan MRM in increasing mobility andfunctional ability. These results support the findings of previous studiesshowing improvement after mobilization in a frozen shoulder.3,12 Additionally, movement strategies interms of scapulohumeral rhythm improved after 3 weeks of MWMtreatment.For the predominant adhesive capsule and associated soft tissue tightness of FSS, mobilization techniqueshave been most commonly addressed in clinical treatment approaches and research studies.3,16 –23Mobilization techniques improve theVolume 87normal extensibility of the shouldercapsule and stretch the tightenedsoft tissues to induce beneficial effects. Our results support thispremise and indicate that the mostbeneficial effects can be achievedwith ERM or MWM, and not MRM,techniques. Although MRM mightextend the adhesive capsule, we believe that the adhesive capsule andassociated contracted periarticularstructures can only be stretched byERM or MWM.Attention to abnormal scapulohumeral rhythm during arm elevation should be increased in rehabilitation programs for subjects withFSS. Vermeulen et al13 observed 10subjects with unilateral FSS for 3months and indicated that improvement in glenohumeral motion following a 3-month period of physicaltherapy intervention did not significantly correspond to normalizationof abnormal scapular motion. Consistent with their findings, our subjectsshowed abnormal scapulohumeralrhythm after 3-month treatments.Normalization of scapulohumeralNumber 10Physical Therapy f1311

Mobilization Techniques for Frozen Shoulder Syndromeafter these alternative treatments, weexcluded these data to avoid biasingour results. Additionally, similar results were found by including dropout data in the intention-to-treat analysis, which further validates ourfindings.Because of substantial FLEX-SF variation of improvement in the relativelysmall sample size between ERM andMWM groups, the lack of statisticalsignificance may have been due totype II error (not enough power).We considered a FLEX-SF score difference of 3 points between groups(minimal clinically important difference and responsiveness were 3.02and 1.12, respectively, for theFLEX-SF in Cook and colleagues’ investigation26) to be clinically meaningful. Using the obtained standarddeviation (5.7) between the 2groups, the power was .38 to detecta FLEX-SF score difference of 3points between groups ( .05). Asample size of 50 subjects per groupwould have been required to achievea power level of .80 to detectFLEX-SF score difference of 3 pointsbetween the 2 groups. Thus, a different treatment effect between ERMand MWM groups is likely and needsto be further investigated.Figure 1.Flow diagram indicating progress of subjects through the study and stage at whichsubjects were lost to follow-up. A mid-range mobilization, B end-range mobilization,C mobilization with movement.rhythm, however, was achievedwith MWM techniques in our subjects. Furthermore, improved mobility and functional ability also wereobserved after MWM treatment.These findings suggest to us thatMWM could increase mobility andimprove motor strategies with regard to the scapulohumeral rhythmin people with FSS.Completion is difficult for subjects ina study that demonstrates no improvement with the intervention.The overall participation rates were1312fPhysical TherapyVolume 87less than in another study,12 wherecompletion rates were 96 out of 116(83%) at 12 months. We recruited 30subjects, of whom 23 (77%) completed the full 12-week study. Themost common reason for droppingout was unwillingness of the subjectto continue due to a lack of improvement following treatment. Five subjects without significant improvement dropped out during MRMtreatment. These subjects were allowed to have alternative treatments(eg, ERM or MWM techniques). Although they showed improvementsNumber 10No benefit was shown during MRMtreatment, but different missing datadue to subjects dropping out due tolack of improvement at 3 and 9weeks between the 2 groups makesinterpretation difficult. We addressed this by secondary analysis(ie, analysis of dropping out between2 groups and survival analysis).There were no differences in numbers of subjects dropping out and nosignificant differences in the survivalexperiences of the 2 groups. Thesefindings suggest that the multipletreatment trial on our 2 groups wasbalanced. It may be, however, thatsubjects continued in the treatmentfor reasons other than treatmenteffectiveness.October 2007

Mobilization Techniques for Frozen Shoulder SyndromeTable 2.Mean Values of Change in Main Outcome Measures in Mobilization Groups and End-Range Mobilization and Mobilization WithMovement Effect Compared With Mid-Range Mobilization Effect After RandomizationaOutcomeMeasureFLEX-SFArm elevation ( )Scapular tipping ( )Mean Changes (95% CI) for A-B-A-C GroupMean Changes (95% CI) for A-C-A-B Mid-RangeMobilizationMobilizationWithMovement5.1 (3.9–6.3)b4.5 (3.1–5.9)b0.2 ( 1.6–1.4)11.7 (5.5–17.9)b6.9 (1.2–11.2)b3.2 ( 5.6–8)0.1 ( 3.9–4.0)0.4 ( 1.9–2.8)1.7 ( 0.3–3.7)Scapulohumeralrhythm0.2 ( 0.1–0.3)Humeral lateralrotation ( )Humeral medialrotation ( )bEnd-RangeMobilizationMid-RangeMobilization7.0 (1.2–13.2)b5.9 (1.2–11.2)b2.3 ( 0.8–6.3)17.6 (9.2–22.1)b6.0 (1.2–11.4)b3.5 ( 2.3–6.8)0.4 ( 3.2–4.0)1.1 ( 0.1–2.4)1.1 ( 3.5–1.3)b0.3 (0.1–0.4)0.1 ( 0.1–0.2)0.2 (0.1–0.3)0.1 ( 0.1–0.2)0.1 ( 0.1–0.2)12.4 (9.1–15.8)b9.1 (6.4–11.8)b3.4 ( 3.5–10.3)7.5 (1.2–10.3)b8.9 (3.2–11.6)b1.1 ( 4.6–5.3)4.1 (0.2–7.9)b2.1 ( 1.3–5.4)1.1 ( 4.4–5.5)4.0 (0.2–8.0)b2.0 ( 1.3–5.5)0.3 ( 5.2–4.7)aA mid-range mobilization, B end-range mobilization, C mobilization with movement, CI confidence interval, FLEX-SF Flexilevel Scale of ShoulderFunction.bP .05.Although our results favored theMWM and ERM treatment techniques, the appropriate treatmentdecision for subjects with FSS maybe dependent on the course and duration of symptoms. Reeves4 documented 3 phases with which to address the progression of FSS: thepain phase, the stiffness phase, andthe recovery phase. Our subjectswere in the second phase, with pri-mary idiopathic FSS and a mean duration of complaints of 20weeks.31,32 The results of this study,therefore, cannot be generalized toother subjects at various stages ofsigns or symptoms or with secondary FSS as a result of diabetes, cardiacproblems, stroke, rheumatoid arthritis, or trauma. It should be noted thatthe outcome of treatment in subjectswith secondary FSS has been docu-mented as less successful.33 Additionally, our multiple-treatment design limits the generalizability of ourfindings to normal clinical practice.Although cumulative effects of mobilizations may be expected at the12-week point, our results at the6-week point (12 visits) are more reasonable for application to normalclinical practice. Additionally, cointervention of MWM and ERM treat-Table 3.Mean Percentage of Change ( SD) in Main Outcome Measures in End-Range Mobilization Effect Compared With MobilizationWith Movement EffectaOutcomeMeasureabMean Percentage of Change at 6 WeeksBetween GroupsMean Percentage of Change at 12 WeeksBetween -SF19.9 8.117.25 12.22.7 ( 5–11)17.9 6.119.2 10.22.2 ( 4–10)Arm elevation ( )11.3 15.18.6 7.85.6 ( 8–10.1)10.3 18.28.8 4.83.6 ( 5–7.1)Scapular tipping ( )31.4 46.318.8 28.412.7 ( 42–68)28.4 46.315.8 29.410.7 ( 40–62)25.7 7.615.9 11.712.8 (4–27)b32.7 21.335.2 12.319.5 21.440.6 32.5Scapulohumeralrhythm10.7 7.624.9 11.7Humeral lateralrotation ( )36.4 24.334.2 14.3Humeral medialrotation ( )20.5 24.445.6 38.5Difference(95% CI)b14.3 (6–22)2.2 ( 16–20)25.3 ( 8–36)Difference(95% CI)3.2 ( 14–18)21.3 ( 5–32)CI confidence interval, FLEX-SF Flexilevel Scale of Shoulder Function.P .05.October 2007Volume 87Number 10Physical Therapy f1313

Mobilization Techniques for Frozen Shoulder SyndromeFigure 2.Summary kinematic data and disability index. A mid-range mobilization, B end-range mobilization, C mobilization with movement, FLEX-SF Flexilevel Scale of Shoulder Function.1314fPhysical TherapyVolume 87Number 10October 2007

Mobilization Techniques for Frozen Shoulder Syndromement techniques may be more beneficial and needs to be furtherinvestigated.Jing-lan Yang, Dr Chang, Dr Wang, and DrLin provided concept/idea/research design.Shiau-yee Chen, Dr Wang, and Dr Lin provided writing. Jing-lan Yang, Shiau-yeeChen, and Dr Lin provided data collection.Shiau-yee Chen and Dr Lin provided dataanalysis. Jing-lan Yang provided projectmanagement and facilities/equipment. DrLin provided fund procurement. Jing-lanYang and Dr Chang provided subjects. DrChang provided institutional liaisons andconsultation (including review of manuscriptbefore submission).This study was approved by the NationalTaiwan University Hospital Review Board.This study was funded by the National Science Council, Taiwan (NSC 94-2314B-002-088).This article was submitted September 27,2006, and was accepted May 22, 2007.DOI: 10.2522/ptj.20060295References1 Neviaser TJ. Adhesive capsulitis. OrthopClin North Am. 1987;18:439 – 443.2 Neviaser RJ

sive mobilization techniques, vary-ing the plane of elevation or varying the degree of rotation in the end-range position, were applied. Mobilization With Movement The use of MWM for peripheral joints was developed by Mulligan.19,20 This technique combines a sustained appli-cation of a manu