FKT

fuKaAbS

B

Preliminary Development of a Clinical Prediction Rule forTreatment of Patients With Suspected SLAP Tears

Stephanie D. Moore-Reed, Ph.D., A.T.C., W. Ben Kibler, M.D.,Aaron D. Sciascia, M.S., A.T.C., and Tim Uhl, Ph.D., P.T., A.T.C.

Purpose: To use the clinical prediction rule process to identify patient variables, measured on initial clinical presentation,that would be predictive of failure to achieve satisfactory improvement, while following a rehabilitation program, in themodification of SLAP injury symptoms and dysfunction.Methods: A cohort of patients received the clinical diagnosis of aSLAP lesion based on specific history and examination findings and/or magnetic resonance imaging. They underwent aphysical examination of the kinetic chain and shoulder, including tests for labral injury. Patients followed a standardizedphysical therapy program emphasizing restoration of demonstrated strength, flexibility, and strength-balance deficits. At6 weeks’ follow-up, patients were re-evaluated and divided into those recommended for surgery (RS) and those notrecommended for surgery (NRS). Bivariate logistic regression was performed to identify the best combination of predictivefactors. Results: Fifty-eight patients (aged 39 � 11 years, 45 men) were included. Of these, 31 (53%) were categorized asNRS and 27 (47%) as RS. The presence of a painful arc of motion (odds ratio, 3.95; P¼ .024) and the presence of increasedforward scapular posture (odds ratio, 1.27; P ¼ .094) on the injured side were predictive of being in the RS group. Thisfinding indicates that the odds of being in the RS group increased 4 times when a positive painful arc was present andincreased 27% with every 1-cm increase in involved anterior shoulder posture. Conclusions: A structured rehabilitationprogram resulted in modification of symptoms and improved function at 6 weeks’ follow-up in over half of patients in thestudy group. On initial evaluation, the presence of a painful arc of overhead motion, indicating loss of normal gleno-humeral kinematics, and the presence of forward shoulder posture, indicating an altered scapular position, representnegative predictive factors for success of rehabilitation. Future validation of the model in a larger population is necessary.Level of Evidence: Level II, prospective comparative study.

here is a growing awareness of the need to develop

Tmore efficacious methods of evaluating, accuratelydiagnosing, and treating all types of shoulder pathology.The clinical prediction rule process has been developed asa method to identify which patient variables, measuredon initial clinical examination, could be predictive of

From the Department of Kinesiology, California State University (S.D.M-R.),resno, California; Shoulder Center of Kentucky (W.B.K., A.D.S.), Lexington,entucky; and Division of Athletic Training, University of Kentucky (A.D.S.,.U.), Lexington, Kentucky, U.S.A.The authors report the following potential conflict of interest or source ofnding: The Lexington Clinic entered into a contract to pay the University ofentucky for a doctoral student, S.D.M-R. (co-author), as a part-time researchssistant as she completed her doctoral degree. T.U. receives support frommerican Society of Shoulder and Elbow Therapists. T.U. received reim-ursement to present components of these data at the International Congress ofhoulder and Elbow Therapists in April 2013 in Nagoya, Japan.Received September 11, 2013; accepted June 13, 2014.Address correspondence to Aaron D. Sciascia, M.S., A.T.C., 1221 Sroadway, Lexington, KY 40504, U.S.A. E-mail: ascia@lexclin.com� 2014 by the Arthroscopy Association of North America0749-8063/13668/$36.00http://dx.doi.org/10.1016/j.arthro.2014.06.015

Arthroscopy: The Journal of Arthroscopic and Related

success of a particular treatment program. The predictionrule process has the potential to instill continuity anduniformity within health care decision making and canhelp guide care.1

This article reports the outcomes of a study applyingthe clinical prediction rule process to a cohort of patientsdetermined to have clinical findings consistent with aSLAP injury. In patients with this diagnosis, surgicaloutcomes are not uniformly successful,2,3 in partbecause of inconsistency in the diagnostic process andinconsistency in surgical indications.4 Rehabilitationprotocols are often advocated as the first step in thetreatment of SLAP and other shoulder injuries, but evi-dence regarding the exact indications, role, and effec-tiveness of rehabilitation in SLAP injury is sparse and notclear. One previous retrospective study reported that49% of patients managed with nonoperative rehabili-tation had a positive outcome.5 Another study, per-formed in professional baseball players, all of whom hada SLAP injury and inwhoma rehabilitation programhadfailed, showed that a comprehensive rehabilitationprogram resulted in 40% of players returning to play.6

Surgery, Vol -, No - (Month), 2014: pp 1-10 1

Fig 1. Participant flowchart.

2 S. D. MOORE-REED ET AL.

The purpose of this prospective study was to use aclinical prediction rule process to attempt to identifypatient variables, measured on initial clinical presenta-tion, that would be predictive of failure to achievesatisfactory improvement, while following a rehabili-tation program, in the modification of symptoms anddysfunction in patients with a clinical diagnosis of SLAPinjury. We hypothesized that in patients who did notrespond, there would be specific measurable patientfactors that were associated with failure to respond to arehabilitation program.

Methods

PatientsFifty-eight patients (mean age, 39 � 11 years; mean

mass, 83 � 25 kg; mean height, 170 � 35 cm) whowere determined to have a clinical history consistentwith shoulder dysfunction resulting from injury to thesuperior labrum were included.7 The history includedpain at the posterior joint line, pain with abduction/external rotation, popping, clicking on shoulder rota-tion, and pain or limitation of performance withrepetitive overhead activity. These 58 patients werepart of a group of 211 patients presenting with shoulderpain to an orthopaedic surgeon (W.B.K.) (Fig 1). Allparticipants were patients of the lead author (W.B.K.),a sports medicine and shoulder surgeon with more than30 years of experience in clinically evaluating andtreating shoulder pathology. There are varying opinionsregarding the exact criteria to establish a diagnosis of aclinically significant SLAP tear, with no literature-established gold standard that is universally recog-nized. Therefore patient inclusion was based on meeting

specific criteria relating to history, clinical examinationfindings, and/or diagnostic imaging findings to includeall possible criteria for detecting the anatomic andfunctional alterations associated with the SLAP injury.The clinical examination inclusion criteria were modi-fied from previously identified criteria for diagnosinglabral tears reported by Walsworth et al.8 The criteria forour study required a positive finding for at least 3 of thefollowing 4 clinical signs: history of popping or catching,positive anterior-slide maneuver, positive modified dy-namic labral shear (M-DLS) maneuver,9 or positiveactive compression test.8,9 Patients with a SLAP teardiagnosed by advanced imaging (i.e., magnetic reso-nance imaging [MRI] or magnetic resonance arthrog-raphy [MRA]) were included if they also had 1 or 2 ofthe clinical examination inclusion criteria because theaddition of history and clinical examination evidencehas been shown to change the information from theimaging.10 Reliance was not placed on 1 examination orimaging test because no single test has been shown to beuniformly satisfactory to make the complete diag-nosis.4,11,12 A recent systematic review by Hegeduset al.13 supports the concept of using clusters of tests tomake the diagnosis in shoulder pathology, although theM-DLS maneuver has been shown in a Level I study tohave high clinical utility.9

Patients were excluded from the study if they hadnumbness or tingling in the upper extremity; signs andsymptoms consistent with cervical radiculopathy,14

adhesive capsulitis,15 or glenohumeral arthritis16;patient-reported steroid injections in the involvedshoulder within the previous month; or surgery on theinvolved shoulder within the past year. They were alsoexcluded if they had clinical examination and/or im-aging findings consistent with a diagnosis of acromio-clavicular joint injury/arthrosis, glenohumeralinstability, or full-thickness rotator cuff tear. This studywas approved by the appropriate institutional reviewboards. Before enrollment in the study, all patients readand signed an informed consent form that wasapproved by the institutional review boards of theUniversity of Kentucky and Lexington Clinic.Patients completed a standard history form and

underwent standard examination by the orthopaedicsurgeon (W.B.K.). All patients completed a numeric painrating scale regarding current pain, worst pain, and leastpain in the past week17 (0, no pain; 10, highest pain). Inaddition, patients completed the Quick Disabilities of theArm, Shoulder and Hand (QuickDASH) questionnaire,which is scored from 0, no disability, to 100, severedisability; the American Shoulder and Elbow Surgeons(ASES) Shoulder Assessment Form, with a score rangingfrom 0, poor function, to 100, normal function18-20; andthe Patient-Specific Functional Scale (PSFS).21 The PSFSquestionnaire requires the patient to list 3 to 5 activitiesthat he or she has difficulty doing because of his or her

LABRAL CLINICAL PREDICTION DEVELOPMENT 3

shoulder problem and to rate each item from 0 (cannotperform activity at all) to 10 (can perform activity at thesame level as before the injury).Glenohumeral range of motion (ROM), strength, and

posture were also assessed. ROM was assessed with adigital inclinometer (Dualer; JTech Medical, Salt LakeCity, UT). Passive internal and external rotation ROMand horizontal adduction ROM were measured withthe patient supine and shoulder abducted to 90� withthe scapula stabilized until resistance was first felt or thepatient reported pain, as previously described.22 Activeshoulder flexion ROM was measured with the patientseated. The patient was instructed to raise his or herarm as high as possible in the sagittal plane with thethumb up.23 The clinician aligned the inclinometer withthe long axis of the humerus, and the angle wasrecorded in degrees. Inter-rater intraclass correlationcoefficients (ICCs) were calculated a priori for internalrotation ROM (ICC, 0.795), external rotation ROM(ICC, 0.839), horizontal adduction ROM (ICC, 0.518),and active flexion ROM (ICC, 0.863).Isometric muscle strength was measured with a

handheld dynamometer24 (model 01163; LafayetteInstruments, Lafayette, IN). Forward flexion strengthwas measured with the participant seated, the scapulain a retracted position, the shoulder in 90� of flexion,and the palm down.24 The dynamometer was placedjust proximal to the wrist, and the participant wasinstructed to push up for 5 seconds. External rotationstrength was measured with the participant supine, theshoulder abducted to 90�, the elbow flexed to 90�, andthe humerus in neutral rotation and supported. Thedynamometer was placed parallel with the forearm. Foreach strength measure, 2 maximum-effort trials wereperformed and averaged for analysis. Each arm wastested in alternating fashion to allow for approximately30 seconds of rest between trials (flexion strength ICC,0.897; external rotation strength ICC, 0.842).Finally, scapular posture was assessed with the

participant standing and using a double square instru-ment as previously described by Kluemper et al.25 Theparticipant was asked to stand against the wall and as-sume his or her normal posture after taking a deepbreath to relax. The double square instrument wasaligned with the wall and the anterior aspect of theacromion. This distance was measured and recordedbilaterally. Reliability was determined a priori (ICC,0.946).After the initial clinical visit and data collection, all

patients were prescribed physical therapy and providedwith a standardized rehabilitation protocol consisting ofstretching exercises and strengthening exercises forshoulder musculature, but the protocol was individu-alized depending on each patient’s examination find-ings by the treating physical therapist. This protocol waswell outlined (Table 1). It consisted of 4 phases, each

having mobility and strengthening components thatwere progressed at each phase. The protocol wasdesigned based on the concepts put forth by Ellen-becker and Cools26 for treating patients with shoulderpain and scapular dysfunction. Mobility exercises pro-gressed from gentle mobility to static stretching ofposterior, anterior, and inferior shoulder mobilityrestrictions. Strengthening exercises progressed fromscapular muscular orientation to gain motor control,using the kinetic chain theories of incorporating theentire body, to short and then to long lever-arm resis-tive exercises, on the basis of kinetic chain theories ofincorporating the lower extremity. Ballistic and eccen-tric exercises were incorporated in the protocol if thetreating therapist believed that they were appropriatefor an individual patient (Table 1). The rehabilitationprotocol was provided to the patients at the initial visit.Patients were allowed to go to the physical therapists oftheir choosing, with instructions to follow the specificprotocol. The physical therapists were provided a letterdescribing the study and requesting that the patientsfollow the established protocol. Exercise logs wereprovided for the patients to record their compliancewith the therapy. Detailed physical therapy recordswere obtained from over two-thirds of the patients(40 of 58).At a follow-up visit with the orthopaedic surgeon

(W.B.K.) 6 weeks after the initial visit (median, 6 weeks;range, 4 to 24 weeks), participants again completed theQuickDASH questionnaire, ASES form, numeric painrating scale, and PSFS questionnaire. Strength, ROM,and posture were also reassessed. In addition, the GlobalRating of Change (GROC) score was obtained. TheGROC is a 15-point scale ranging from �7 (a great dealworse) to þ7 (a great deal better), with 0 indicating nochange.27 Exercise logs and physical therapy notes werecollected from patients at this time.After the intervention and follow-up appointment,

patients were categorized into 2 groups based on theirreport of their clinical status and the clinical examina-tion findings: recommended for surgery (RS) or notrecommended for surgery (NRS). The recommendationfor surgery was based on continued or worsened sub-jective and objective symptoms of shoulder pain anddysfunction, failure to progress in rehabilitation, and apatient’s unwillingness or inability to tolerate thedysfunction, with clinical input from the physician(Table 2). This process followed the normal procedureof consultation and decision making regarding treat-ment between physician and patient. The decision tocounsel and recommend surgery was made at this timepoint because most studies indicate that it takes around6 weeks to observe significant changes in physiologicalfactors such as flexibility and strength and thereforeaffect the clinical symptoms, which were the goals ofthe rehabilitation protocol.28-30 In addition, this is

Table 1. Rehabilitation Exercise Program

Exercise Category Level I Level II Level III Level IV

Scapular orientation Scapular protraction andretraction (e.g., scapular clock)

Scapular and humeral depression(e.g., inferior glide)

Muscle strengtheningScapular retraction Below shoulder level: isometric Below shoulder level: isotonic (e.g.,

dynamic low row, lawnmower,robbery)

Isotonic at shoulder level: short leverarm (e.g., pull downs, fencing, rows)

Isotonic (e.g., prone horizontal abductionlifts at 90� or 135�)

Scapular protraction Punch (e.g., supine punch, scapularpunches)

Push-ups (e.g., incline) Push-ups (e.g., knee, standard)

Punch (e.g., standing punch) Diagonal (e.g., upper cut)Humeral rotation Below shoulder level: isotonic

(e.g., IR/ER with arm at side withresistance)

At shoulder level (e.g., ER/IR with elasticband, 90�/90�)

Humeral rotation at shoulder level (e.g.,90�/90� rotation, side-lying ER eccentricexercises)

Humeral elevation Short lever arm (e.g., overhead press) Long lever arm (e.g., flexion, abduction,plyometrics, weighted-ball drops)

StretchingAnterior Supine pectoral stretch with arm

at sideSupine pectoral stretch with

overpressureActive scapular retraction with arms at 90�

ER with arm at side ER with arm away from sidePosterior Cross body Sleeper stretch Sleeper stretch in more abducted positionElevation Table slides Wall slides Latissimus dorsi stretch Active latissimus dorsi stretch

Forward bows Assisted elevation with pulley

ER, external rotation; IR, internal rotation.

4S.

D.MOORE-REED

ETAL.

Table 2. Improvements in Outcome Measures From InitialEvaluation to 6 Weeks’ Follow-up

Outcome MeasureNRS Group(n ¼ 31)

RS Group(n ¼ 27) P Value

ASES score 14 � 16 2 � 16 .006QuickDASH score �13 � 15 0 � 16 .001Current pain �1 � 2 0 � 2 .047PSFS score 2 � 3 0 � 3 .033GROC score 3 � 2 0 � 2 <.001

NOTE. Data are change scores and are given as mean � SD unlessotherwise indicated.

Table 4. Counts and Percentages for Categorical Variables

VariableNRS Group(n ¼ 31)

RS Group(n ¼ 27)

Apprehension 2 (6%) 2 (7%)Anterior load and shift 1 (3%) 0 (0%)Posterior jerk 1 (3%) 0 (0%)Sulcus 1 (3%) 2 (7%)ER lag 0 (0%) 0 (0%)Drop arm 0 (0%) 0 (0%)Belly press 0 (0%) 0 (0%)Liftoff 4 (13%) 0 (0%)Bear hug 6 (19%) 5 (19%)Upper cut 11 (35%) 11 (41%)Speed’s test 5 (16%) 6 (22%)M-DLS 28 (90%) 22 (81%)Active compression test 18 (58%) 14 (52%)Anterior slide 21 (68%) 11 (41%)Point-tender pain* 8 (26%) 3 (11%)Crepitus 3 (10%) 2 (7%)Paxinos test 3 (10%) 3 (11%)Pain with horizontal adduction 3 (10%) 4 (15%)Painful arc (forward flexion)y 15 (48%) 21 (78%)Hawkins impingement 10 (32%) 11 (41%)Neer impingement 9 (29%) 9 (33%)SICK scapular position 18 (58%) 18 (67%)Scapular dyskinesis 30 (97%) 23 (85%)Scapular assistance test* 17 (55%) 20 (74%)Scapular retraction test 21 (68%) 18 (67%)Single-leg balance* 13 (42%) 7 (26%)Single-leg squat 12 (39%) 9 (33%)Pain with resisted abduction 26 (84%) 27 (100%)Patient-reported pop/grind/click 24 (77%) 20 (74%)Previous physical therapy* 12 (39%) 14 (52%)Gradual onset of symptoms 17 (55%) 14 (52%)Male sex 23 (74%) 22 (81%)

ER, external rotation; SICK, scapular malposition inferior medialborder prominence coracoid pain and malposition and dyskinesis of

LABRAL CLINICAL PREDICTION DEVELOPMENT 5

standard procedure for the physician’s practice. Thepatient-reported outcome measures were also analyzedand compared with the RS/NRS status.

Statistical AnalysisDescriptive statistics were calculated for each variable

measured at baseline for continuous (Table 3) andcategorical (Table 4) data. Variable selection was per-formed by testing each variable’s univariate associationwith the outcome by use of Student independent t testsfor continuous variables and c2 tests for categoricalvariables. Variables with a univariate P � .30 wereretained, leaving 5 continuous variables (scapularposture, ASES score, QuickDASH score, PSFS score, andcurrent pain) and 5 categorical variables (scapularassistance test, painful arc test, previous physical ther-apy, point-tender pain, and single-leg balance) as po-tential predictors in the model. Backward stepwiselogistic regression was used to determine the relativecontribution of each variable to outcome (i.e., RS orNRS). Five variables were entered into the final

Table 3. Descriptive Statistics for Continuous Variables

VariableNRS Group(n ¼ 31)

RS Group(n ¼ 27)

Height (cm) 175 � 10 178 � 8Mass (kg) 85 � 21 89 � 18Age (yr) 40 � 11 39 � 55Duration of symptoms (mo) 19 � 11 25 � 42ROM (�)

Flexion 144 � 28 142 � 22ER 75 � 24 76 � 28IR 60 � 17 61 � 20HAdd �10 � 10 �9 � 9

Strength (lb)ER 10 � 6 9 � 5Flexion 7 � 3 8 � 3

Scapular posture (cm)* 14 � 2 15 � 2ASES score* 62 � 16 56 � 21QuickDASH score* 34 � 16 41 � 17Current painy 3.5 � 1.8 4.7 � 2.5PSFS score* 3.6 � 1.9 3.1 � 1.7

NOTE. Data are given as mean � SD.ER, external rotation; HAdd, horizontal adduction; IR, internal

rotation.*Significant at P � .30.ySignificant at P � .05.

scapular movement.*Significant at P � .30.ySignificant at P � .05.

backward stepwise logistical regression analysis (scap-ular posture, scapular assistance test, painful arc test,previous physical therapy, and point-tender pain).Compliance was examined in 2 ways: average num-

ber of physical therapy visits per week and compliancewith the rehabilitation protocol. Compliance with theprotocol was determined by the percentage of exercisesperformed by the patient that were from the protocolprovided to the physical therapists as part of studyparticipation. Student independent t tests were per-formed for average visits per week and compliance withthe protocol. Statistical analyses were conducted usingSPSS software, version 19 (SPSS, Chicago, IL).

ResultsOf the 58 patients, 27 (47%) were categorized as

RS at 6 weeks. Twenty-six of these 27 patients hadadvanced imaging performed; in 24 of these, the im-aging yielded positive findings for a SLAP injury.Twenty-two of the 27 patients in the RS group (81%)

Table 5. Logistic Regression Predicting “Recommended forSurgery”

VariableRegressionCoefficient SE P Value OR (95% CI)

Scapularprotraction

0.242 0.145 .094 1.27 (0.96 to 1.69)

Painful arc test 1.375 0.609 .024 3.95 (1.20 to 13.05)Constant �4.581 2.141 .042

CI, confidence interval; OR, odds ratio; SE, standard error.

6 S. D. MOORE-REED ET AL.

went on to undergo arthroscopic surgery, and all werefound to have an anatomic labral injury that met thecriteria for a clinically significant SLAP injury.11 The 5patients who did not undergo surgery ultimatelydeclined because of insurance issues or did not schedulesurgery after the recommendation.Thirty-one patients (53%) were categorized as NRS at

6 weeks. Eighteen of these 31 had advanced imagingperformed; in 16 of these, the imaging yielded positivefindings for a SLAP injury. Two patients in the NRSgroup (6%) eventually had enough functional limita-tions that they returned after the study and elected toundergo surgery. Both patients were found to have anarthroscopically confirmed anatomic labral injury thatmet the criteria for a clinically significant SLAP injury.11

The logistic regression analysis showed that thepresence of a painful arc of motion in forward flexionand the presence of increased forward scapular postureon the injured side at the initial examination werepredictive of being in the RS group at 6 weeks (P ¼.015, df ¼ 2, c2 ¼ 8.413) (Table 5). The odds of being inthe RS group increased 4 times when a positive painfularc was present and increased 27% with every 1-cmincrease in involved anterior shoulder posture. Thefinal model predicted 72% of the outcomes correctly.Detailed physical therapy records were collected and

analyzed for 19 of 27 patients in the RS group and 21 of31 patients in the NRS group. There was no significantdifference in average visits per week between the RSpatients (1 � 2 weeks) and NRS patients (2 � 1) (P ¼.638). There was also no significant difference betweencompliance with the provided rehabilitation protocolbetween the RS patients (55% � 38%) and NRS pa-tients (71% � 29%) (P ¼ .159).There was a statistically significant difference in 3

patient-reported outcome measuresdASES score,QuickDASH score, and GROC scoredbetween the RSand NRS groups. The NRS group reported statisticallysignificantly larger improvements in these scales, whichwould provide some further confirmation for the basisfor the treatment decision (Table 1).

DiscussionThis study confirms the research hypothesis; the

clinical prediction rule process identified 2 specific

variables, a painful arc of motion in forward flexionand increased anterior scapular position, as factorsassociated with failure to achieve a satisfactoryimprovement so that the patient and physician agreedthat surgery was recommended. A structured reha-bilitation program resulted in modification of symp-toms, increased outcome measures, and improvedfunction to the point that the patient and physicianagreed that surgery was not necessary in 31 patients(53%) in the study group at 6 weeks, although 2 ofthese patients ended up requesting surgery at a latertime. Because these patients were not recommendedto undergo surgery at the 6-week time point of in-terest, they were kept in their original group (RS orNRS) as determined by the physician at 6 weeks forfollow-up analysis.The clinical prediction rule process has the potential

to be a significant element in improving the delineationof the factors that can influence the content andtiming of treatment of musculoskeletal injury. Clinicalprediction models serve as formal, evidence-basedapproaches to clinical decision making by using statis-tical models to provide quantitative estimates of prob-ability of outcome, diagnosis, or treatment success.31-33

They have the potential to instill continuity and uni-formity within health care decisions and can help guidecare.1,31 However, the validity and clinical impact of aprediction rule must be determined before the model istranslated to a clinical decision rule intended to affectclinical decision making.34 The need for more carefullyderived and validated prediction models in orthopae-dics and rehabilitation has been identified.35 Thesemodels have the potential to help the patient byavoiding surgery while continuing to restore acceptablefunction, help the physician by more clearly definingwhich patients need surgery, and help the health caresystem by increasing the efficiency of the content andtiming of treatment. This particular model was designedto try to identify patient factors that could influence theoutcome of a specific treatment program (a rehabilita-tion program) in patients with clinical findings sugges-tive of a SLAP injury.The research decision to evaluate this process in this

population was based on several factors. SLAP lesionsare being more commonly diagnosed and treated. SLAPinjuries are reported to be present in 6% to 12% ofshoulder arthroscopies,36-38 and the incidence of sur-gery appears to be increasing.2,39 The rationale foroperative intervention is to restore the anatomic alter-ation in the labrum and its attachment to the glenoidthat is assumed to be responsible for the patient’s painand dysfunction. However, surgical outcomes varywidely and are not uniformly successful,40-43 high-lighting the need for a more consistent diagnostic pro-cess and for a better understanding of more preciseindications for and timing of surgery.

LABRAL CLINICAL PREDICTION DEVELOPMENT 7

Attention toward using rehabilitation as the first stepin the treatment of patients with SLAP lesions hasincreased. Protocol content would be directed towardimproving motion and strength deficits, maximizingkinetic chain function, and modifying or minimizingthe dysfunction that these patients report. There are afew reports that show the benefits of this approach. Arecent retrospective study observed that 19 of 39 pa-tients (49%) diagnosed with SLAP lesions treatednonoperatively reported successful outcomes at 3 yearsafter diagnosis and that 10 of 15 athletes (67%) hadreturned to their preinjury status.5 The exact rehabili-tation program was not described completely. Anotherstudy examined treatment of SLAP lesions in profes-sional baseball players.6 These players had undergone 1session of rehabilitation that had failed. A specific pro-gram of nonoperative intervention focused on correct-ing scapular dyskinesis and glenohumeral internalrotation deficit resulted in a 40% rate of return to play,although not all of these patients returned to theirpreinjury level of performance.The specific rehabilitation program used in our study

was designed using the concepts put forth by Ellen-becker and Cools26 and is similar to previous pro-tocols.6,11 This study adds further support to the use ofthis type of rehabilitation program as the first method oftreatment because both studies report a similar successrate. These results show that not all patients whoreceive the diagnosis of a SLAP injury will requiresurgery.This study used statistical analysis to identify variables

associated with a particular outcome. It was notdesigned to determine why or by what mechanismsthese variables were associated with the outcome.However, the findings were consistent with the ideathat clinically symptomatic and dysfunctional SLAP le-sions represent not only anatomic disruption of thelabrum and its glenoid attachment but also local anddistant physiological and biomechanical alterations thatadd up to create the clinical dysfunction.11,44 None ofthe variables most commonly associated with makingthe diagnosis of an anatomic labral injury (clinical testssuch as ROM deficits,44 apprehension in externalrotation or positive biceps stress testing,45 a positive M-DLS maneuver,9 an active compression test,46 or posi-tive MRI findings) were associated with failure oftherapy and the need for surgery. This finding suggeststhat the presence of physical impairments, in additionto the results of isolated clinical maneuvers or imaging,may provide stronger clinical relevance for makingtreatment decisions.This study used the clinical prediction rule process to

derive an actual clinical prediction rule that can helpguide management of patients presenting with theaforementioned symptoms. The next step wouldinvolve validation of this rule with a prospective study

in a different cohort of patients with the same clinicalpresentation. One existing clinical prediction model hasbeen developed47 and externally validated48 to predictpersistent symptoms at 6 weeks in patients withshoulder pain seen in the primary care setting. At6 weeks, Kuijpers et al.47 observed that a longer dura-tion of symptoms, gradual onset of pain, psychologicalcomplaints, report of repetitive movements at least2 days per week, and high pain severity in the shoulder(scale from 0 to 10) and in the neck (scale from 0 to 18)at presentation were associated with persistent symp-toms. One fundamental difference between the existingmodel and our analysis is that our patients were pre-scribed a standardized rehabilitation protocol whereasthe existing model was developed for patients treatedprimarily with medication, corticosteroid injection, or a“wait-and-see” approach. In addition, our study spe-cifically included patients who had symptoms consis-tent with a SLAP lesion and all patients were seen by anexperienced orthopaedic surgeon (W.B.K.).A significant percentage of patients with clinical

findings suggestive of SLAP lesions that have createdshoulder dysfunction can show modification of symp-toms with lessening or elimination of the dysfunctionthrough a specific rehabilitation program and do notrequire or request surgery. Subjective reports ofimproved function can be objectively documentedusing valid and reliable patient-reported outcomemeasures. This finding is in agreement with andreinforces the limited information in the literature.However, 2 patient variables found on clinical exami-nation, a positive painful arc of motion in forwardflexion between 60� and 100� and forward scapularposture, were found to be associated with failure tosucceed in the therapy program. The odds of rehabili-tation failure increased 3.95 times with the presence ofa painful arc of motion and increased 27% with every1-cm increase in forward scapular position.Factors that have been traditionally used to delineate

the anatomic SLAP lesion, such as isolated clinical ex-amination tests, glenohumeral rotation asymmetries,responses to scapular corrective maneuvers, or MRI, didnot predict success or failure of rehabilitation. Reha-bilitation can be advocated as the first step in thetreatment of most patients with SLAP tears. Rehabili-tation can frequently modify the physiological alter-ations that are a large part of the shoulder dysfunction,even in the face of an anatomic labral injury, andprovide restoration of function so that surgery is notdeemed necessary by the patient and the clinician. Atthis time, patients who, on initial examination, have apainful arc of motion and scapular protraction and whoundergo rehabilitation have higher odds of eventuallybeing recommended for surgery. These 2 factors mayindicate a greater alteration of glenohumeral kine-matics associated with the clinical dysfunction. Future

8 S. D. MOORE-REED ET AL.

directions from this study would include a validationstudy to determine the effectiveness of the clinicalprediction rule in larger populations or in anotherpopulation with these findings, as well as extension toother shoulder pathology.

LimitationsThere are several limitations to this study. The first

and most important is that the diagnosis of a “labralinjury” was made by a combination of clinical exami-nation and imaging techniques, with no actual visual-ization of the lesion. However, there is no accepted goldstandard for making this diagnosis either clinically, byimaging, or by arthroscopy.10,13,49 Most studies evalu-ating clinical examination tests have found poor speci-ficity for the detection of actual anatomic labralinjuries.13 The M-DLS test appears to have the highestsensitivity, specificity, and likelihood ratio when per-formed correctly9 but is still best used as part of acomprehensive evaluation.11 MRI has high sensitivitybut unreliable specificity for clinically significant labralinjury, as shown by the high prevalence of changesconsistent with labral injury seen in asymptomaticoverhead athletes.50 Even direct arthroscopic visuali-zation has shown poor reliability in diagnosing andclassifying labral injuries.4 In the absence of a recog-nized gold standard, current practice guidelines werereviewed to establish consistent criteria to distinguishthis group of patients with suspected SLAP tears.In the absence of specific clinical examination tests,

specific imaging criteria, or even arthroscopic criteria toestablish a single standard, many recent authorsrecommend basing the diagnosis on clusters of exam-ination and imaging findings.9,10,12,13 For our cohort, itwas decided to use a modification of the criteria pro-posed by Walsworth et al.8 for the clustering of clinicaltests. Patients had to have 3 or more history and ex-amination findings to be included. All patients whohad fewer than 3 of the history and examination in-clusion criteria but had a diagnosis of labral injury bymagnetic resonance arthrogram were also includedbecause most studies use this as an inclusion criterion.The exact percentages for each of the criteria are listedin Table 4. Because very few patients in the NRS groupunderwent arthroscopic evaluation, it may be thatsome patients in this group did not actually have ananatomic SLAP injury. However, all patients in bothgroups had the same cluster of inclusion criteria andexhibited the same clinical dysfunction. Our consis-tently applied inclusion criteria and the documentedexclusion criteria provided a fairly homogeneousgroup whose dysfunction has usually been consideredto be due to SLAP injury.11,51 This study group isconsidered to have clinical findings suggestive of aSLAP injury6,11 and can be differentiated from othergroups with other diagnoses of shoulder injury.

The efficacy of this diagnostic approach was strength-ened by the arthroscopic findings compatible with aclinically significant SLAP injury7 in every patient whounderwent surgery.The second limitation relates to the lack of control

over the rehabilitation program performed by the pa-tients. However, the physical therapy notes showedthat both groups of patients participated in a high per-centage of the expected sessions and completed theexercises as prescribed, and the available therapists’notes showed progression in the protocols. This type ofrehabilitation management is consistent with the real-world scenario of how therapy is performed andimplemented, and the overall impression was thatcompliance was slightly higher than usual.The third limitation is that the decision for surgery

was determined after only 6 weeks of therapy, andperhaps a longer time of follow-up could havebetter delineated the response. However, Tate et al.52

observed that the greatest rate of improvement in theDisabilities of the Arm, Shoulder and Hand question-naire occurred in the first 2 weeks of treatment.Because the decision for surgery was based on thesurgeon’s findings of no significant improvement onthe clinical examination or in symptoms, as well as thepatient’s experience of no significant change in thedysfunction, we believe that this decision represents anaccurate representation of the effects of the rehabilita-tion program. The accuracy of this determination at thistime is substantiated by the percentages of patients ineach group who eventually required surgery.The fourth limitation is that our patient group rep-

resents an active and recreationally athletic populationbut not a population of professional athletes. Theapplicability of these results to the professional athletemay be limited because of the higher level of demands.This study should be viewed as a preliminary report. It

shows that the clinical prediction process can develop amodel that elucidates factors that are associated withthe results of an intervention. The model needs to bevalidated in a larger population. Finally, the resultsoccurred with the use of a specific rehabilitation pro-tocol. There may be other protocols that may showimproved results.

ConclusionsA structured rehabilitation program resulted in

modification of symptoms and improved function at6 weeks’ follow-up in over half of patients in the studygroup with clinical dysfunction and symptoms sugges-tive of a SLAP injury. On initial evaluation, the pres-ence of a painful arc of overhead motion, indicatingpotential loss of normal glenohumeral kinematics, andthe presence of forward shoulder posture, indicatinga potentially altered scapular position, represent nega-tive predictive factors for success of rehabilitation.

LABRAL CLINICAL PREDICTION DEVELOPMENT 9

Future validation of the model in a larger population isnecessary.

AcknowledgmentThe authors thank the Physical Therapists of PT Pros

and Lexington Clinic Physical Therapy for helpingcreate the rehabilitation protocol and Kelley Seekins,M.S., A.T.C., for her assistance with data collection.

References1. Rosenberg W, Donald A. Evidence based medicine: An

approach to clinical problem-solving. BMJ 1995;310:1122-1126.

2. Weber SC, Martin DF, Seiler JG III, et al. Superior labrumanterior and posterior lesions of the shoulder: Incidencerates, complications, and outcomes as reported byAmerican Board of Orthopaedic Surgery part II candi-dates. Am J Sports Med 2012;40:1538-1543.

3. Brockmeier SF, Voos JE, Williams RJ III, et al. Outcomesafter arthroscopic repair of type-II SLAP lesions. J BoneJoint Surg Am 2009;91:1595-1603.

4. Gobezie R, Zurakowski D, Lavery K, et al. Analysis ofinterobserver and intraobserver variability in the diag-nosis and treatment of SLAP tears using the Snyder clas-sification. Am J Sports Med 2008;36:1373-1379.

5. Edwards SL, Lee JA, Bell JE, et al. Nonoperative treat-ment of superior labrum anterior posterior tears:Improvements in pain, function, and quality of life. Am JSports Med 2010;38:1456-1461.

6. Fedoriw WW, Ramkumar P, McCulloch PC, et al. Returnto play after treatment of superior labral tears in profes-sional baseball players. Am J Sports Med 2014;42:1155-1160.

7. Kibler WB. What is a clinically important superior labrumanterior to posterior tear? Instr Course Lect 2013;62:483-489.

8. Walsworth MK, Doukas WC, Murphy KP, et al. Reliabilityand diagnostic accuracy of history and physical exami-nation for diagnosing glenoid labral tears. Am J Sports Med2008;36:162-168.

9. Kibler WB, Sciascia AD, Hester P, et al. Clinical utility oftraditional and new tests in the diagnosis of biceps tendoninjuries and superior labrum anterior and posterior lesionsin the shoulder. Am J Sports Med 2009;37:1840-1847.

10. Wolf BR, Britton CL, Vasconcellos DA, et al. Agreement inthe classification and treatment of the superior labrum.Am J Sports Med 2011;39:2588-2594.

11. Kibler W, Kuhn J, Wilk K, et al. The disabled throwingshoulder: Spectrum of pathologyd10-year update.Arthroscopy 2013;29:141-161.

12. McFarland EG, Kim TK, Savino RM. Clinical assessmentof three common tests for superior labral anterior-posterior lesions. Am J Sports Med 2002;30:810-815.

13. Hegedus EJ, Goode AP, Cook CE, et al. Which physicalexamination tests provide clinicians with the most valuewhen examining the shoulder? Update of a systematicreview with meta-analysis of individual tests. Br J SportsMed 2012;46:964-978.

14. Wainner RS, Fritz JM, Irrgang JJ, et al. Reliability anddiagnostic accuracy of the clinical examination and pa-tient self-report measures for cervical radiculopathy. Spine2003;28:52-62.

15. Griggs SM, Ahn A, Green A. Idiopathic adhesive cap-sulitis. A prospective functional outcome study ofnonoperative treatment. J Bone Joint Surg Am 2000;82:1398-1407.

16. Kelley MJ, Ramsey ML. Osteoarthritis and traumaticarthritis of the shoulder. J Hand Ther 2000;13:148-162.

17. Childs JD, Piva SR, Fritz JM. Responsiveness of thenumeric pain rating scale in patients with low back pain.Spine 2005;30:1331-1334.

18. Gummesson C, Ward MM, Atroshi I. The shortened dis-abilities of the arm, shoulder and hand questionnaire(QuickDASH): Validity and reliability based on responseswithin the full-length DASH. BMC Musculoskelet Disord2006;7:44.

19. Beaton DE, Wright JG, Katz JN. Development of theQuickDASH: Comparison of three item-reduction ap-proaches. J Bone Joint Surg Am 2005;87:1038-1046.

20. Michener LA, McClure PW, Sennett BJ. AmericanShoulder and Elbow Surgeons Standardized ShoulderAssessment Form, patient self-report section: Reliability,validity, and responsiveness. J Shoulder Elbow Surg2002;11:587-594.

21. Stratford P, Gill C, Westaway M, et al. Assessing disabilityand change on individual patients: A report of a patientspecific measure. Physiother Can 1995;47:258-263.

22. Moore SD, Laudner KG, McLoda TA, et al. The immediateeffects of muscle energy technique on posterior shouldertightness: A randomized controlled trial. J Orthop SportsPhys Ther 2011;41:400-407.

23. Valentine RE, Lewis JS. Intraobserver reliability of 4physiologic movements of the shoulder in subjects withand without symptoms. Arch Phys Med Rehabil 2006;87:1242-1249.

24. Hayes K, Walton JR, Szomor ZL, et al. Reliability of 3methods for assessing shoulder strength. J Shoulder ElbowSurg 2002;11:33-39.

25. Kluemper M, Uhl TL, Hazelrigg H. Effect of stretching andstrengthening shoulder muscles on forward shoulderposture in competitive swimmers. J Sport Rehabil 2006;15:58-70.

26. Ellenbecker TS, Cools A. Rehabilitation of shoulderimpingement syndrome and rotator cuff injuries:An evidence-based review. Br J Sports Med 2010;44:319-327.

27. Jaeschke R, Singer J, Guyatt GH. Measurement of healthstatus. Ascertaining the minimal clinically important dif-ference. Control Clin Trials 1989;10:407-415.

28. Sale DG. Neural adaption to resistance training. Med SciSports Exerc 1988;20:S135-S145 (suppl).

29. Bandy WD, Irion JM, Briggler M. The effect of time andfrequency of static stretching on flexibility of thehamstring muscles. Phys Ther 1997;77:1090-1096.

30. Moritani T, deVries HA. Neural factors versus hypertro-phy in the time course of muscle strength gain. Am J PhysMed 1979;58:115-130.

31. Grady D, Berkowitz SA. Why is a good clinical predictionrule so hard to find? Arch Intern Med 2011;171:1701-1702.

10 S. D. MOORE-REED ET AL.

32. Beattie P, Nelson R. Clinical prediction rules: What arethey and what do they tell us? Aust J Physiother 2006;52:157-163.

33. Braitman LE, Davidoff F. Predicting clinical states inindividual patients. Ann Intern Med 1996;125:406-412.

34. Reilly BM, Evans AT. Translating clinical research intoclinical practice: Impact of using prediction rules to makedecisions. Ann Intern Med 2006;144:201-209.

35. Simmen BR, Bachmann LM, Drerup S, et al. Develop-ment of a predictive model for estimating the probabilityof treatment success one year after total shoulderreplacementdCohort study. Osteoarthritis Cartilage2008;16:631-634.

36. Snyder SJ, Banas MP, Karzel RP. An analysis of 140injuries to the superior glenoid labrum. J Shoulder ElbowSurg 1995;4:243-248.

37. Handelberg F, Willems S, Shahabpour M, et al. SLAPlesions: A retrospective multicenter study. Arthroscopy1998;14:856-862.

38. Maffet MW, Gartsman GM, Moseley B. Superior labrum-biceps tendon complex lesions of the shoulder. Am J SportsMed 1995;23:93-98.

39. Onyekwelu I, Khatib O, Zuckerman JD, et al. The risingincidence of arthroscopic superior labrum anterior andposterior (SLAP) repairs. J Shoulder Elbow Surg 2012;21:728-731.

40. Yung PS, Fong DT, Kong MF, et al. Arthroscopic repair ofisolated type II superior labrum anterior-posterior lesion.Knee Surg Sports Traumatol Arthrosc 2008;16:1151-1157.

41. Neuman BJ, Boisvert CB, Reiter B, et al. Results ofarthroscopic repair of type II superior labral anteriorposterior lesions in overhead athletes: Assessment ofreturn to preinjury playing level and satisfaction. Am JSports Med 2011;39:1883-1888.

42. Cohen DB, Coleman S, Drakos MC, et al. Outcomes ofisolated type II SLAP lesions treated with arthroscopicfixation using a bioabsorbable tack. Arthroscopy 2006;22:136-142.

43. Kim SH, Ha KI, Choi HJ. Results of arthroscopic treatmentof superior labral lesions. J Bone Joint Surg Am 2002;84:981-985.

44. Burkhart SS, Morgan CD, Kibler WB. The disabledthrowing shoulder: Spectrum of pathology Part I: Patho-anatomy and biomechanics. Arthroscopy 2003;19:404-420.

45. Burkhart SS, Morgan CD, Kibler WB. Shoulder injuries inoverhead athletes, the “dead arm” revisited. Clin SportsMed 2000;19:125-158.

46. O’Brien SJ, Pagnani MJ, Fealy S, et al. The activecompression test: A new and effective test for diagnosinglabral tears and acromioclavicular joint abnormality. Am JSports Med 1998;26:610-613.

47. Kuijpers T, van der Windt DA, Boeke AJ, et al. Clinicalprediction rules for the prognosis of shoulder pain ingeneral practice. Pain 2006;120:276-285.

48. Kuijpers T, van der Heijden GJ, Vergouwe Y, et al. Goodgeneralizability of a prediction rule for prediction ofpersistent shoulder pain in the short term. J Clin Epidemiol2007;60:947-953.

49. Phillips JC, Cook C, Beaty S, et al. Validity of noncontrastmagnetic resonance imaging in diagnosing superiorlabrum anterior-posterior tears. J Shoulder Elbow Surg2013;22:3-8.

50. Han K-J, Kim Y-K, Lim S-K, et al. The effect of physicalcharacteristics and field position on the shoulder andelbow injuries of 490 baseball players: Confirmation ofdiagnosis by magnetic resonance imaging. Clin J Sport Med2009;19:271-276.

51. Provencher MT, McCormick F, Dewing C, et al.A prospective analysis of 179 type 2 superior labrumanterior and posterior repairs: Outcomes and factorsassociated with success and failure. Am J Sports Med2013;41:880-886.

52. Tate AR, McClure PW, Young IA, et al. Comprehensiveimpairment-based exercise and manual therapy interven-tion for patients with subacromial impingement syndrome:A case series. J Orthop Sports Phys Ther 2010;40:474-493.