THE CLINICAL COURSE OF shoulder pain is commonly. Conservative Treatment for Shoulder Pain: Prognostic Indicators of Outcome

Transcription

1 1231 Conservative Treatment for Shoulder Pain: Prognostic Indicators of Outcome Karen A. Ginn, PhD, Milton L. Cohen, MD ABSTRACT. Ginn KA, Cohen ML. Conservative treatment for shoulder pain: prognostic indicators of outcome. Arch Phys Med Rehabil 2004;85: Objectives: To investigate the long-term clinical outcome and to identify factors that predict that outcome, after conservative treatment of patients who have shoulder pain with or without accompanying stiffness. Design: Cohort study. Setting: Outpatient clinic. Participants: Eighty-two subjects who had participated in a randomized controlled trial that compared the short-term effectiveness of conservative treatment for chronic, unilateral shoulder pain of mechanical origin with and without accompanying stiffness, and who were available for longer term follow-up 6 months after the cessation of formal treatment. Interventions: Conservative treatment consisting of various combinations of exercise therapy, passive joint mobilization, electrophysical modalities, and corticosteroid injections. Main Outcome Measures: Pain intensity, functional limitation, perceived change in symptoms, active range of motion, muscle force, and clinical and demographic variables. Results: Patients showed significant improvement in all outcome measurements in the long-term whether or not their shoulder pain was accompanied by stiffness. Long-term outcome was not predicted by hand dominance, clinical history of the shoulder condition, severity of the shoulder problem, or shoulder mechanics. Conclusions: Patients with chronic shoulder pain, with or without accompanying stiffness, can expect significant decreases in shoulder pain and improvements in shoulder function in the long term after conservative treatment. Key Words: Injections, intra-articular; Physical therapy techniques; Prognosis; Rehabilitation; Shoulder pain; Treatment outcome by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation THE CLINICAL COURSE OF shoulder pain is commonly chronic. Between 22% and 46% of patients who visit a medical practitioner because of shoulder pain report a history of pain 1,2 ; one survey 2 reported an average of 6 previous episodes. Six months after initial consultation and treatment, persistent shoulder symptoms have been reported in 34% to 79% of patients, 1,3 with between 24% and 61% of patients still From the Faculty of Health Sciences, University of Sydney, Lidcombe (Ginn); and Faculty of Medicine, University of New South Wales, Sydney (Cohen), Australia. Supported by the University of Sydney (grant nos. S342, P1504, S ). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Karen Ginn, PhD, Sch of Biomedical Sciences, Faculty of Health Sciences, University of Sydney, PO Box 170, Lidcombe 1825, Australia, K.Ginn@fhs.usyd.edu.au /04/ $30.00/0 doi: /j.apmr reporting pain 6 to 18 months later. 1,3-5 Of those with persistent symptoms, more than half typically do not seek additional treatment. 4 In addition, surveys 6,7 indicate that only about half of older people in the community who have persistent shoulder pain seek any medical attention. Various factors have been associated with the outcome of treatment for shoulder pain. More rapid recovery has been associated with a shorter duration of symptoms, 1,3 a history of overuse associated with the shoulder pain, 5 and after slight trauma or an unusual activity. 4 By contrast, patients with shoulder pain who initially present with concomitant neck pain, 4 have severe pain, 4 severe functional impairment, 3 muscle weakness, 3 or decreased passive range of motion 1 (PROM), or who have had previous shoulder pain, 1 were more likely to have persistent or recurring symptoms after treatment. Our purpose in this study was to investigate long-term clinical outcomes after conservative treatment of patients with shoulder pain with or without accompanying stiffness and to identify factors that predict the outcomes in those patients. METHODS This study was approved by the Ethics Committees of the University of Sydney and St Vincent s Hospital, Sydney, Australia. Participants Our subjects comprised a subgroup of volunteers who had participated in a clinical trial 8 that compared the short-term effectiveness of conservative treatments for shoulder pain and who were available for reassessment 6 months after their formal treatment ended. These subjects had pain over 1 shoulder joint and/or into the proximal arm that was exacerbated by active shoulder movements. Subjects with pain in both shoulders, or pain associated with shoulder instability or with inflammatory or neoplastic disorders, pain referred from vertebral column structures, or pain resulting from recent trauma, were not included. The study cohort included 82 subjects who had received various forms of conservative treatment for shoulder pain. Of the 82, 47 had decreased abduction and/or flexion range of motion (ROM) accompanying the shoulder pain (PR subgroup) and 35 had a painful arc of abduction and/or flexion motion, but retained full ROM (P subgroup). Procedure Eligible subjects signed a consent form before undergoing a standardized interview about the medical history of their shoulder condition, including the duration of the current symptoms, factors that may have precipitated the onset of the symptoms, hand dominance, and previous episodes of shoulder pain. A musculoskeletal assessment was made to obtain baseline outcome measurements of pain intensity, functional limitation, ROM, and muscle force. Reliable measurements of shoulder mechanics were also made. These included curvature of the upper thoracic spine, position of the scapula at rest, and range and force of shoulder internal and external rotation. Subjects had been randomly allocated to receive 1 of 3 conservative treatments for 5 weeks: exercises aimed at restor-

2 1232 SHOULDER PAIN AND LONG-TERM TREATMENT OUTCOME, Ginn ing neuromuscular control at the shoulder; treatment consisting of electrophysical modalities, passive joint mobilization, and ROM exercises (multiple modalities treatment); or a subacromial, corticosteroid injection. The short-term effectiveness of the treatment approaches we used has been confirmed in randomized controlled trials of acceptable methodologic quality If symptoms persisted after the initial 5 weeks of treatment, subjects were offered additional treatment either to continue their originally allocated treatment or to another treatment. Outcome measurements were repeated 6 months after formal treatment ended by independent assessors who were unaware of the subjects treatment histories. Three assessors were involved in this study. Table 1: ICCs for Interrater Reliability for Descriptive Measurements Descriptive Measurements Interrater ICC 2,1 Thoracic curve.53 Horizontal scapular alignment.77 Lateral scapular slide.58 Internal rotation ROM.74 External rotation ROM.87 Internal rotation force.69 External rotation force.86 Outcome Measurements Pain intensity was measured on a 10-cm vertical visual analog scale labeled no pain and severe pain at its extremes, immediately after performance of a standardized reaching task. Functional limitation associated with the shoulder pain was measured by using an individually standardized, self-reported score. Each subject was asked to rate the level of difficulty in performing 9 specified upper-limb tasks often affected by shoulder dysfunction. Subjects were also asked to identify any other tasks that particularly provoked shoulder symptoms. A 4-point scale of increasing difficulty (range: 0, can perform with no shoulder pain; 3, cannot perform because of shoulder pain) was used. Only the tasks subjects identified at the initial interview as causing shoulder pain were rated at the follow-up assessment. A functional limitation score was obtained by summing the scores for each item; the higher the score, the greater the functional limitation associated with the shoulder pain. Perceived change in symptoms was measured at reassessment by a self-report from the subject on a 3-point scale that included getting better, staying the same, and getting worse. Active abduction and flexion ROM were measured by using the photographic method we have previously described. 13 For subjects in the PR subgroup, the point in range where pain was first felt, termed onset of pain, and the limit of ROM, termed ROM, were measured. For subjects in the P subgroup, the painful excursion, termed painful ROM, was measured. Hand-behind-back range was determined by measuring the distance between T1 spinous process and the radial styloid process with a tape measure. A hand-behind-back ROM score was obtained by subtracting the affected side measurement from the unaffected side. This relative measurement, using the asymptomatic shoulder as the control, was chosen because normative data for hand-behind-back range was not available and clinical experience has indicated that there is great variation in hand-behind-back ROM in the unimpaired population. Isometric abduction force was measured with a hand-held dynamometer that has shown acceptable reliability when tested with patients with strength deficits. 13,16 Isometric abduction force for the affected side was recorded as a percentage of the unaffected side. Descriptive Measurements Curvature of the upper thoracic spine was measured by using a plurimeter and was defined as the difference between the inclination of the spine at T1 and T7 spinous processes. Curvature of the upper thoracic spine is related proportionally to trunk height; therefore, upper thoracic spine curvature measurements were standardized to the subject s sitting height before analysis. 17 Photography was used to measure the horizontal plane alignment of the scapula. The posterolateral aspects of both acromion processes were marked with the subject standing, feet shoulder-width apart, and arms relaxed by the side. Horizontal alignment was determined by the angle between the line passing through the marks on the acromion processes and a horizontal reference line passing through the posterolateral acromion process of the unaffected side. Positive angle values were used to indicate that the affected side was higher than the unaffected side. Lateral scapular slide (distance of the scapula from the midline) was determined by measuring the horizontal distance from the tip of the T4 spinous process to the medial border of the scapula with a tape measure. A lateral scapular slide score was obtained by subtracting the affected side measurement from the unaffected side measurement. Photography was also used to measure active shoulder internal and external rotation ROM. Subjects were in crook lying, with their shoulder abducted to 90 (or to 45 of abduction if the subject could not achieve 90 ) and elbow flexed to 90. The olecranon and ulnar styloid processes were marked on the subject while he/she was in this position. Internal and external rotation ranges were determined bilaterally from the angle between a line passing through the olecranon and the ulnar styloid processes and a horizontal reference line. Active shoulder internal and external rotation ROM was recorded as a percentage of the unaffected side. With subjects in the supine position, isometric internal and external rotation forces were measured bilaterally using a handheld dynamometer, with the shoulder at 90 of abduction (or to 45 of abduction if the subject could not achieve 90 ). The arm was fully supported, elbow flexed to 90, and the forearm pronated. The dynamometer was placed just proximal to the ulnar styloid process. A make test was used. The subject was asked to build to a maximum contraction over a 1- to 2-second period and then to hold for a further 4 to 5 seconds. On the affected side, the force at which pain was first felt was recorded and the maximum force produced was recorded for the unaffected side. Isometric rotation force for the affected side was recorded as a percentage of the unaffected side. The intraclass correlation coefficient (ICC 2,1 ) 18 showed good to excellent reliability for all these descriptive measurements (table 1). Data Analysis Perceived change in symptoms, expressed as a percentage and means and medians, including 95% confidence intervals (CIs), for pain intensity, functional limitation, ROM, and isometric abduction force measurements were calculated before treatment began and 6 months after formal treatment ceased. Median changes in pain intensity, functional limitation, all ROM measurements, and abduction force before treatment and 6 months after treatment were compared by using Wilcoxon signed-rank tests.

3 SHOULDER PAIN AND LONG-TERM TREATMENT OUTCOME, Ginn 1233 Fig 1. Treatment history. Logistic regression analysis was performed by using variables commonly associated with shoulder pain, or previously identified as predictive of treatment outcome for shoulder disorder, that is, age, duration of symptoms, dominant side affected, previous shoulder pain, precipitating factors, pain intensity, severity of functional limitation, and muscle force. 1,3,4,19 In our analysis, a good outcome was defined as a score of 0 for pain intensity, of less than 2 for functional limitation, and subjects describing their condition as better at reassessment 6 months after the treatment ended. Hierarchical logistic regression analysis was performed by using variables that are associated with shoulder mechanics and considered by some clinicians to influence treatment outcome. To standardize 1 of these variables (upper thoracic spine curvature) to sitting height, sitting height was entered as step 1 with upper thoracic spine curvature as step 2, followed by the remaining 6 variables (horizontal plane alignment of the scapula, lateral scapular slide, hand-behind-back ROM, shoulder internal and external rotation ROM, shoulder rotation force ratio) as step 3. A good outcome was defined in the same manner as for the logistic regression analysis described earlier. All analyses were performed by using SPSS software. a Statistical significance was set at P less than.05. RESULTS The treatment history of our 82 subjects is summarized in figure 1. The average time to follow-up after formal treatment ceased was 9 months (range, 7 22mo). Thirty-one subjects (38%) completed formal treatment after 5 weeks. Sixty-five subjects (79%) had been exposed to an exercise program during this clinical trial: 59 subjects (72%) had participated in the exercise treatment program aimed at restoring neuromuscular control at the shoulder, with 23 of these subjects receiving this treatment only; the remaining 6 subjects (7%) had participated only in the ROM exercise component of the multiple modalities intervention (fig 1). The results for all outcome measurements before treatment and 6 months after treatment ceased are represented in table 2. Sixty of the 82 subjects (73%) completing reassessment at 6 months had no pain on the reaching task used to measure pain intensity. Comparison of the median changes between values before treatment and 6 months after treatment ended indicated that subjects improved significantly over the long term in all outcome measurements (table 3). Logistic regression analysis found that duration of symptoms, dominant side affected, previous shoulder pain, precipitating factors, pain intensity, functional limitation, and muscle force did not significantly predict outcome in this study ( , P.08). Age was the only variable having a significant effect on outcome (P.01). Logistic regression analysis using this variable only indicated that age did significantly predict outcome ( , P.00), but the classification table yielded less than impressive results (table 4). Hierarchical logistic regression analysis by using thoracic curve standardized to sitting height, scapula position, handbehind-back ROM, shoulder internal and external rotation ROM, and shoulder rotation force ratio indicated that these factors did not significantly predict outcome (table 5).

4 1234 SHOULDER PAIN AND LONG-TERM TREATMENT OUTCOME, Ginn Table 2: Outcome Measurements Before Treatment and 6 Months After Treatment Ended Outcome Measurements Before Treatment After Treatment Change in symptoms Improving 64 (78%) Stable 10 (12%) Deteriorating 8 (10%) Pain intensity (cm) 1.8 ( ) 0 (0 0) Functional limitation 7 (6 10) 1 (1 3) HBB ROM (cm) (aff unaff)* 6 (3.5 13) 2 (1 6) Abduction force (% unaff) 53 (40 70) 70 (62 100) Painful abduction ROM: P group (deg) 47 (33 59) 1 (0 17) Painful flexion ROM: P group (deg) 22 (4 34) 0 (0 1) Abduction onset of pain: PR group (deg) 64 (57 72) 104 (93 116) Abduction ROM: PR group (deg) 87 (70 92) 145 ( ) Flexion onset of pain: PR group (deg) 94 (76 100) 117 ( ) Flexion ROM: PR group (deg) 107 (98 116) 130 ( ) NOTE. All values are medians and 95% CIs except onset of pain in abduction (mean, 95% CIs) and change in symptoms (n, percentages). Abbreviation: HBB, hand-behind-back. *Aff unaff is the HBB ROM score obtained by subtracting affected side measurement from unaffected side measurement. % unaff is expressed as a percentage of the unaffected side. DISCUSSION Our subjects showed significant improvement in all outcome measurements 6 months after their last treatment. The magnitude of these changes, calculated as a percentage of the pretreatment value of each outcome measure, indicates clinically significant improvement. Subjects reported an 88% improvement in functional capacity, improvements of from 63% to 98% in pain-free and/or total abduction and hand-behind-back ROM, and 32% improvement in abduction force measured as a percentage of the unaffected side (table 2). In addition, 78% of subjects reported that their shoulder problems had improved during the trial (table 2). Although median shoulder pain levels had improved significantly, 22 subjects (27%) were still reporting some shoulder pain at reassessment 6 months after their final treatment. Of these, 4 subjects (5%) were reporting decreased pain, 10 subjects (12%) reported no change in their shoulder pain and 8 subjects (10%) were experiencing increased pain (table 2). This rate of persistence of pain after treatment falls at the lower end of reported chronicity rates for shoulder pain. 1,3-5 Subjects in this clinical trial, therefore, achieved long-term success rates comparable with the best estimates reported in the literature. This study did not identify any factors associated with the onset, duration, or initial severity of the subject s shoulder dysfunction that significantly influenced the outcome of treatment in the long term. Specifically, severity of pain, functional Table 3: Median Changes Between Outcome Measurements Before Treatment and 6 Months After Treatment Ended Outcome Measurement Z Scores (P Values) Pain intensity 5.99 (.000) Functional limitation 7.25 (.000) HBB ROM 5.17 (.000) Abduction force 4.14 (.000) Abduction onset of pain (PR group) 4.03 (.000) Abduction ROM (PR group) 5.27 (.000) Flexion onset of pain (PR group) 3.39 (.001) Flexion ROM (PR group) 4.76 (.000) Painful abduction (P group) 3.46 (.001) Painful flexion (P group) 3.06 (.002) limitation, restriction of movement, or degree of muscle force deficit were not predictive of treatment outcome. In addition, neither a history of previous shoulder pain nor the resting positions of the scapula or upper thoracic vertebral column at entry into the study were associated with long-term treatment success. Although direct comparison between studies is difficult because of variation in the manner in which treatment success and outcomes were measured, these results support the findings of Bartolozzi et al, 3 who reported no effect of arm dominance or active ROM on the clinical outcome after conservative treatment for shoulder pain. In contrast, previous reports of a poor outcome after conservative treatment associated with the duration of symptoms before treatment, 1,3 increased initial severity of symptoms, 4 severe functional limitation before treatment, 1,3 hand dominance, 5 and previous episodes of shoulder pain 1 were not supported by the results of this study. The only characteristic that significantly predicted outcome in this study was the age of the subject at the time of recruitment to the trial: the older the subject, the less the likelihood he/she was to have a good outcome. It might be argued that because degenerative changes in shoulder joint structures are common in the elderly (eg, glenoid labrum lesions, 20 rotator cuff tears ), this result may have been expected. However, many elderly people with significant structural abnormality do not suffer from shoulder pain or limitation of function. For example, rotator cuff lesions have been shown in 50% of asymptomatic subjects more than 60 years of age 21,23 and in approximately 80% of symptom-free subjects more than 80 years of age. 21 In this study, the age of the subject correctly predicted a poor outcome in less than 50% of subjects (table 4). In other words, more than half of the older subjects had a successful response to treatment in the long term. There are several reasons for factors predictive of long-term treatment outcome failing to be identified in this study. This Table 4: Classification Table for Age Predicted Not Good Predicted Good % Correct Observed not good % Observed good %

5 SHOULDER PAIN AND LONG-TERM TREATMENT OUTCOME, Ginn 1235 Table 5: Hierarchical Logistic Regression Results Step 1 (sitting height) Step 2 (thoracic curve) Step 3 (all other variables) , P , P , P.49 study only reports the long-term outcome in 82 subjects, therefore, the logistic regression analysis may have lacked the power to detect significant differences. The size of this cohort, however, was sufficient to detect clinically significant differences in all outcome measurements, from entry to long-term follow-up. Second, because the self-assessed improvement scale used did not include completely recovered or much improved categories, treatment success was described by a composite measure that included a score of getting better on the self-assessed improvement scale, no pain on the standardized reaching task, and minimal functional impairment from shoulder pain. This modest definition of treatment success may have contributed to our not identifying factors that influence treatment outcome. In addition, we did not investigate some of the presenting factors reported elsewhere to be predictive of treatment outcome, for example, concomitant neck pain 4 and decreased PROM. 1 On the other hand, some aspects of the treatment given in this trial may have influenced the prognosis for participating subjects. The exercise intervention, aimed at restoring neuromuscular control of the shoulder, and the multiple modalities intervention were individually determined for each subject by the treating physical therapist, based on the musculoskeletal assessments. Perhaps this patient-specific approach to treatment is more likely to address the particular mechanical problems associated with a person s shoulder dysfunction and thus result in a better treatment outcome for that person. Such a conclusion would concur with the opinion of the well-known orthopedic physician, Cyriax, who described the shoulder joint as the most rewarding joint in the body. It passes the salient merits of honesty and curability. When some movement is found painful or limited, the significance is seldom equivocal... (in Daigneault and Cooney 24(p1145) ). CONCLUSIONS Our results indicate that, after conservative treatment, patients with chronic shoulder pain with or without accompanying stiffness can expect a significant decrease in shoulder pain and improvement in shoulder function in the long term. However, this outcome is not predicted by hand dominance, clinical history of shoulder pain, initial severity of the shoulder condition or shoulder resting posture, shoulder ROM, or shoulder muscle force. Acknowledgments: We thank the physical therapy staff at St Vincents Hospital, Sydney, Australia, for their willing assistance in this study; Dr. Robert Heard (Faculty of Health Sciences, University of Sydney Australia) for his help in the analysis of these data; and Dr. Roslyn Bohringer (Faculty of Health Sciences, University of Sydney Australia) for her advice during preparation of this manuscript. References 1. Croft P, Pope D, Silman A. The clinical course of shoulder pain: prospective cohort study in primary care. BMJ 1996;313: van der Windt DA, Koes BW, de Jong BA, Bouter LM. Shoulder disorders in general practice: incidence, patient characteristics, and management. Ann Rheum Dis 1995;54: Bartolozzi A, Andreychik D, Ahmad S. Determinants of outcome in the treatment of rotator cuff disease. Clin Orthop 1994; Nov(308): van der Windt DA, Koes BW, Boeke AJ, Deville W, de Jong BA, Bouter LM. Shoulder disorders in general practice: prognostic indicators of outcome. Br J Gen Pract 1996;46: Chard M, Sattelle L, Hazleman B. The long term outcome of rotator cuff tendinitis a review study. Br J Rheumatol 1988;27: Chakravarty K, Webley M. Shoulder joint movement and its relationship to disability in the elderly. J Rheumatol 1993;20: Chard M, Hazleman R, Hazleman B, King R, Reiss B. Shoulder disorders in the elderly: a community survey. Arthritis Rheum 1991;34: Ginn K. Exercise therapy aimed at restoring neuromuscular control for the treatment of shoulder pain: a comparative clinical trial [dissertation]. Sydney: Univ New South Wales; Adebajo AO, Nash P, Hazleman BL. A prospective double blind dummy placebo controlled study comparing triamcinolone hexacetonide injection with oral diclofenac 50mg TDS in patients with rotator cuff tendinitis. J Rheumatol 1990;17: Bang M, Deyle G. Comparison of supervised exercise with and without manual physical therapy for patients with shoulder impingement syndrome. J Orthop Sports Phys Ther 2000;30: Conroy D, Hayes K. The effect of joint mobilization as a component of comprehensive treatment for primary shoulder impingement syndrome. J Orthop Sports Phys Ther 1998;28: Dacre JE, Beeney N, Scott DL. Injections and physiotherapy for the painful stiff shoulder. Ann Rheum Dis 1989;48: Ginn K, Herbert R, Khouw W, Lee R. A randomized, controlled clinical trial of a treatment for shoulder pain. Phys Ther 1997;77: Petri M, Dobrow R, Neiman R, Whiting-O Keefe Q, Seaman WE. Randomized, double-blind, placebo-controlled study of the treatment of the painful shoulder. Arthritis Rheum 1987;30: Winters J, Sobel J, Groenier K, Arendzen H, Meyboom-de Jong B. Comparison of physiotherapy, manipulation, and corticosteroid injection for treating shoulder complaints in general practice: randomised, single blind study. BMJ 1997;314: Bohannon R. Testing isometric limb strength with dynamometers. Clin Rev Phys Rehabil Med 1990;2: Raine S, Twomey L. Posture of the head, shoulders and thoracic spine in comfortable erect standing. Aust J Physiother 1994;40: Shrout P, Fleiss J. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979;86: Chard M, Hazleman B, Devereux M. Controlled trial to investigate dose-response patterns to portable pulsed electromagnetic field in the treatment of rotator cuff tendinitis. J Orthop Rheumatol 1988;1: Andrews J, Carson W, McLeod W. Glenoid labrum tears related to the long head of biceps. Am J Sports Med 1985;13: Milgrom C, Schaffler M, Gilbert S, van Holsbeeck M. Rotatorcuff changes in asymptomatic adults. The effect of age, hand dominance and gender. J Bone Joint Surg Br 1995;77: Petersson C. Degeneration of the gleno-humeral joint: an anatomical study. Acta Orthop Scand 1983;54: Sher JS, Uribe JW, Posada A, Murphy BJ, Zlatkin MB. Abnormal findings on magnetic resonance images of asymptomatic shoulders. J Bone Joint Surg Am 1995;77: Daigneault J, Cooney L. Shoulder pain in older people. J Am Geriatr Soc 1998;46: Supplier a. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL