End Feel 4a71m

Journal of Orthopaedic & Sports Physical Therapy 2OOO;3O (9):5 12-527

Construct Validity of Cyriax's Selective Tension ~xaminahon:A'ssociation of End-feels With Pain at the Knee and Shoulder Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Cheryl M. Petersen, P7; MS Karen W Hayes, P7; PhD

Study Design: Descriptive. Objectives: To examine the relationship between pain and normal and abnormal-pathologic end-feels during ive physiologic motion assessment at the knee and shoulder. We theorized that abnormal-pathologic end-feels would be more painful than normal end-feels. Background: End-feel testing and pain intensity information are part of physical therapy musculoskeletal patient examinations. End-feels are categorized as normal or abnormalpathologic. No previous studies have examined the relationship between pain during endfeel testing and the type of end-feel. Methods and Measures: Two physical therapists examined subjects with unilateral knee or shoulder pain. Each subject was examined twice. ive physiologic motions, 2 at the knee and 5 at the shoulder, were tested by applying an overpressure at the end of range of motion using standardized positions. Subjects reported the amount of pain (0-10) immediately after the evaluator recorded the end-feel. Analyses included one-way ANOVAs and post-hoc Tukey's Honestly Significant Difference tests. Results: Some abnormal-pathologic end-feels were significantly more painful than the normal end-feels at both the knee and the shoulder for all physiologic motions. Among the abnormal-pathologic end-feel categories there were no statistical differences in pain intensity, although small samples in some categories may be responsible for this finding. Conclusion: Abnormal-pathologic end-feels are associated with more pain than normal endfeels during ive physiologic motion testing at the knee or shoulder. Dysfunction should be suspected when abnormal-pathologic end-feels are present. ) Orthop Sports Phys Ther 2000;30:512-527.

Key Words: manual therapy, orthopedics, physical therapy, tests and measurements

Department of Physical Therapy and Human Movement Sciences, Northwestern University Medical School, Chicago, 111. Study ed by the Foundation for Physical Therapy and approved by the lnstitutional Review Board of Northwestern University, Chicago, 111. Send correspondence to: Cheryl M. Petersen, 645 N. Michigan Avenue, Suite 1100, Chicago, I1 606 1 1. E-mail: [email protected]

yriax's system of selective tension testing, used as part of a patient's physical examination, is designed to identify the specific anatomical structure causing the patient's symptoms. Selective tension testing involves using active motion, ive motion, resisted contractions, and palpation to identify soft tissue lesions in inert (ligament, capsule, bursa, fascia, dura mater, and nerve) or contractile tissues (structures that form part of a muscle) by reproducing the patient's complaint, which is often pain, and demonstrating dysfunction in the soft t i s s ~ e s . ~ ( p p ~ ~ - ~ ~ ) Part of Cyriax's selective tension theory involves the concept of end-feel testing during ive movements. The different sensations that the examiner perceives at the end of each ive movement are end-feels. Cyriax suggested that end-feels are important in patient management because they either indicate pathology or guide intervention (ie, a bony end-feel should not be manipulated) .J(p5J) End-feels can be normal or abnormal-pathologic, depending on the movement that they accompany at

C

TABLE 1. End-feel classification systems.

blfenbor~~'~~"

CyriaXJW-Yl

Normal

1. Capsular' 2. Bone-to-bone* 3. Tissue approximation

1. Soft 2. Firm 3. Hard

Abnormal-Rthologic

1. Capsular' early in range 2. Bone-to-bone* 4. Spasm 5. Springy block 6. Empty

An end-feel "that occurs at another place or is of another quality than is characteristic for the t being tested."

pa+W-' 1. Soft tissue approximation 2. Muscular 3. Ligamentous 4. Cartilaginous 5. Capsular 6. Capsular (chronidacute) 7. Adhesions and scarring 8. Bony block 9. Bony grate 10. Springy rebound 11. Rnnus 12. Loose 13. Empty 14. Rinful 15. Muscle

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

* Capsular and bone-to-bone end-feels can be normal or abnormal-pathologic, depending on the motion and the point in the range at which they occur. a particular t or the point in the expected range of motion at which the end-feel occurs. Each t has a characteristic normal end-feel, which is dependent on the anatomy of the t and the direction of the physiologic movement tested. Other end-feels at each t would be considered abnormal-pathologic. The few studies on end-feel testing have provided little for the reliability of end-feel testing, or for the theoretical constructs underlying the concept of end-feel, and have provided no strategies to improve end-feel testing. Because additional research is needed, this report is part of a study that examined the components of Cyriax's selective tension testing using Cyriax's end-feel classification system. Pain information obtained at the time of end-feel testing allowed us to study the magnitude of the subject's pain response for normal versus abnormal-pathologic endfeels. We hypothesized that abnormal-pathologic endfeels would occur in a pathologically involved t, and the pain associated with an abnormal-pathologic end-feel would be greater than the pain found with a normal end-feel. The basis for our hypotheses was that pathology at any t suggests changes in the normal physiology of the tissues (inert or contractile) surrounding the t complex. Our goal was to determine whether pain found during end-feel testing was greater with abnormal-pathologic as compared to normal end-feels. If so, clinicians would then be able to identify t regions where pathology was present when they found abnormal-pathologic end-feels. Three different classifications of end-feels are found in the literature. Paris classifies end-feels into groups of 5 normal and 10 abnormal-pathologic endfeel~,'~(p*~)while Cyriax groups end-feels into 6 specific categories with 2 of the end-feels described as being either normal or abnormal-pathologi~."pp~ Both ~ ~ the ) Paris and Cyriax classifications are based on the anatomical structure that limits

range of motion. Kaltenborn, in comparison, uses the nature of the resistance "soft," "firm," or "hard" to describe end-feels. He states that "all three types of end-feel have an elastic quality to varying degrees . . . dependent on the anatomy of the t.""^^*) All 3 systems are summarized in Table 1. In Cyriax's descriptions of his 6 end-feel categories, he states that bone-to-bone, capsular at the end of normal range, and cissue approximation end-feels occur in healthy ts and are normal for some motions. A cap sular end-feel "consists of the hardish arrest of movement, with some give in it, as if two pieces of tough rubber were being squeezed together" such as occurs at the end of shoulder movements. T i e approximation occurs when "the t cannot be pushed farther because of engagement against another part of the body" as at the end of knee and elbow flexion. A bone-to-bone end-feel "is the abrupt halt to movement when two hard surfaces meet," as when bone engages bone during elbow exten~ion.~(F@-) Cyriax describes abnormal-pathologic end-feels as capsular, before normal full range is reached (sometimes called early capsular), spasm, springy block, and empty. An early capsular end-feel is again the "hardish arrest of movement, with some give in it" occurring at the end of the patient's limited range of motion. The "vibrant twang" of the spasm end-feel, as Cyriax described it, relates to the reflex muscle spasm occurring with inflammation. A springy block end-feel occurs when "a rebound is seen and felt at the extreme of the possible range" in ts in which an inua-articular displacement can occur such as in the knee, sternoclaviculart, and spine. When the examiner feels no resistance and there is "considerable pain before the extreme of range is reached," the end-feel is called "empty," and important disease such as acute bursitis, extra-articular abscess, or neoplasm is s ~ g g e s t e d . ~ ( pWe p ~considered ~~) a bone-tobone end-feel abnormal-pathologic when it occurs at a t in which that end-feel is not normally present.

J Orthop Sports Phys Ther .Volume SO. Number 9 September 2000

513

TABLE 2. Characteristicsof subjects. Subject characteristic Age (years)

Mean = 31.8, SD = 9.5, range = 22-60

Sex Side of involvement Highest grade completed in school (years) Duration of condition (months)

18 men, 22 women 20 right, 20 left Mean = 16.0, SD = 1.6, range = 12-1 8 Mean = 33.9, SD = 52.1, range = 0.4-234, median = 7.7 Mean = 173.7, SD = 8.4, range = 157.5-1 88.0 Mean = 76.1, SD = 16.7, range = 44.3-1 13.3

Height (cm) Weight (kg)

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Knee problems (n = 40)

Shoulder problems (n = 46) Mean = 34.3, SD = 12.91, range = 21-75 21 men, 25 women 33 right, 13 left Mean 16.1, SD = 1.9, range = 8-18 Mean = 38.9, SD = 49.9, range = 0.33-243, median = 17.6 Mean = 170.8, SD = 10.3, range = 152.Cb193.0 Mean = 72.5, SD = 13.5, range = 54.3-1 13.3

Because pain is inherent in the names and definitions and is taught in many physical therapist educations of 2 of Cyriax's abnormal-pathologic end-feels, tion programs. The reliability and construct validity of spasm and empty, we expected that these end-feels either the total system or its parts have only recently would be painful when present. Theoretically, when been s t ~ d i e d . ' " ~ J ~Part J ~ of examining construct vaan early capsular, a springy block, or bone-to-bone lidity involves testing hypotheses that should be s u p end-feel occur within a t, pathology exists, and ported if tests perform as predicted from theory. there is reason to believe that these end-feels would We theorized that subjects who had early capsular, be painful as well. According to Cyriax, an early c a p bone-to bone, spasm, springy block, and empty endsular end-feel suggests fibrosis of the capsule or ligafeels at the knee and shoulder would have more pain ments associated with that t. A springy block end- during end-feel testing than subjects with normal feel indicates involvement of a meniscus or disc with- end-feels because these abnormal-pathologic endin the t."pm) A bone-to-bone end-feel indicates feels are purported to suggest some pathologic actividegeneration of articular cartilage within the t. ty in t related structures. We have not seen this The mechanics of a t with an early capsular, idea ed or refuted in the applied experimenspringy block, or bone-to-bone end-feel would be tal literature. The purpose of this report is to examchanged, resulting in probable stress to other tissues ine the hypothesis that during ive physiologic of that t. Pain could be associated with the abmotion testing, subjects with abnormal-pathologic normal stress.I0 In our experience, patients often reend-feels at the knee or shoulder would have greater port pain when abnormal-pathologic end-feels are self-reported pain than subjects with normal endpresent. Cyriax stated that pain associated with "arfeels. thritis" suggests capsular involvement and that early capsular or spasm end-feels could o c c ~ r . ~ ( p p ~ ~ , ~ METHODS ~) There are few studies examining the reliability or usefulness of end-feel testing. Patla and Paris examSubjects ined the reliability of end-feel assessments based on Paris' classification system at the elbow for 20 asymp Subjects for the study were recruited from the unitomatic subjects.I2 Inter-rater reliability for 2 examin- versity community, comprising primarily students ers, reported as kappa coefficients, was 0.40 for flexfrom the professional schools and patients from phyion and 0.73 for extension. Intra-rater reliability resician referrals. They all had unilateral knee or ported as percentage of agreement, ranged from 75 shoulder pain. to 80%.12Because asymptomatic subjects were examSubjects with knee pain (18 men and 22 women) ined in this study, the kappa coefficients may have had problems equally distributed between the right been inflated due to the examiner's expectation of and left sides. Subjects with shoulder pain (21 men normal end-feel findings. Hayes et a17 reported intra- and 25 women) had mostly right-sided problems. rater kappa coefficients of 0.17 for knee extension The majority of both groups were in their 20's and and 0.48 for knee flexion end-feels based on Cyriax's 30's; many of them were university graduate students. classification system in subjects with osteoarthritis of For both groups of subjects, the duration of symp the knee. The 2-month time frame between tests in toms was highly variable (Table 2). The median duthis study could have allowed the symptoms to ration was nearly 8 months for subjects with knee change, decreasing reliability. pain and nearly 18 months for subjects with shoulder The basic postulates of the Cyriax selective tension pain. system need to be examined because the system of seA questionnaire and screening physical examinalective tension techniques, including end-feel testing, tion were used to try to exclude subjects with muscuis part of physical therapy musculoskeletal examinaloskeletal or neurological conditions that prevented J Orthop Sports Phys Ther-Volume SO. Number 9.September 2000

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

active motion at the knee and shoulder or pain arising from the spine and viscera. Subjects were included if they had voluntary movement of the knee and shoulder and asymptomatic active and ive movements with overpressures (additional force applied after resistance to motion is felt) of the cervical or lumbar spine. Exclusion criteria included current or previous spinal problems; fractures or dislocations; genitourinary or lower gastrointestinal problems that could refer pain to the knee; gastric, hepatic, respiratory or cardiac problems that might refer pain to the shoulder; multiple sclerosis, cerebrovascular accident, peripheral neuropathies, or other neurological conditions. Inclusion and exclusion criteria were included in subject recruitment materials. As part of the screening process, subjects signed an informed consent form approved by the Institutional Review Board at Northwestern University and the rights of the subjects were protected through the review board's guidelines. No subjects were excluded from the study after the screening examination.

Evaluators A research assistant screened all subjects. She had 12 years of experience, 10 of which were in o r t h e paedic practice, and she continued to practice as a physical therapist during the 4year course of the study. The 2 evaluators, both physical therapists, had studied selective tension testing, and used the system in their practices, but they were not considered experts in Cyriax techniques. The evaluators had 18 and 20 years of experience with at least 10 years of orthopaedic emphasis and continued to practice in orthopaedic physical therapy during the course of the study. The evaluators spent 6 hours with the authors as part of the study reviewing data forms, test procedures, and their interpretation. We collected data based on the way typical practitioners had been taught in professional or continuing education programs, and the evaluators did not practice examining patients together.

ive Tests Performed Only the ive physiologic motion tests, which are part of a knee or shoulder evaluation recommended by Cyriax,were assessed in this study. To test end-feels, the evaluators used standardized stabilization techniques with overpressures for knee flexion and extension and shoulder full abduction, glenohumeral abduction, external rotation, internal rotation, and horizontal adduction. Knee flexion and extension end-feels were evaluated in the supine p o s i t i ~ n . ~ (Knee p ~ ~ )flexion was performed by ively flexing the subject's hip and the knee with an overpressure at the end of knee flexion J Orthop Sports Php Ther.Volume SOeNumber 9-September 2 0 0

range. Knee extension was produced with the hip in a neutral position with an overpressure at the end of knee extension ra~~ge.'~(pp*~*) The expected end-feel without pathology for knee flexion is tissue approximation, and for knee extension, a normal end-feel is capsular. All shoulder end-feels were evaluated in the standing position."(ppflWn3)The expected normal end-feel for all shoulder motions is capsular but may also be tissue approximation for shoulder horizontal adduction. During shoulder movements, specific stabilization of the scapula was used except during full a b duction. During glenohumeral abduction, evaluators stabilized the scapula with their palms over the acromion, applying a force in an inferior direction. Cradling the subject's upper extremity with the elbow flexed, they performed shoulder abduction in the plane of the scapula. To test full shoulder abduction, evaluators provided no scapular stabilization and moved the shoulder through full ive abduction in the frontal plane. During external rotation, evaluators stabilized the subject's scapula with their bodies and performed external rotation by grasping the subject's flexed forearm, keeping the elbow against the subject's body. During internal rotation, evaluators stabilized the scapula in an inferior direction with their fingers over the coracoid process and their palms and forearms over the scapula. They grasped the subject's flexed lower forearm and performed internal rotation. During horizontal adduction, evaluators stabilized the lateral border of the scapula with their palm over the scapula and reached in front of the subject to cradle the flexed upper extremity and produce horizontal a d d u c t i ~ n . ' ~ ( ~ ~ ~ - ~ ~ . ~ ~ )

Procedures The evaluators did not interview the subjects during the examination. We did not want patient interview information to affect the interpretation of the physical examination. The evaluators were told only in which knee or shoulder the pain was located. The subjects were to indicate to the evaluator only if the specific pain complaint was recreated during testing. Evaluators recorded all test results on a data form. Subjects were examined a second time following a lsminute rest period. Subjects were reexamined by the same evaluator (intra-rater condition, n = 23 knee and n = 28 shoulder) or the other evaluator (inter-rater condition, n = 17 knee and n = 18 shoulder). The intra-rater versus inter-rater conditions were randomized, and in the inter-rater situation, the order of evaluator was also randomized. Subjects were instructed to report to the evaluators as soon as they felt the onset of, or an increase in, their pain. Immediately after the evaluator recorded the end-feel, the subject reported the amount of pain he or she experienced. Patients verbally indicat-

TABLE 3. An example of a contingency table for intra-rater ive knee flexion end-feels. (One evaluator completing repeated assessments on 23 patients.) Test Retest

Caps

E Caps

TA

Caps E Caps TA SB S E

5 1

1 1

1

SB

S

E

1

9 1 2 1

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Kappa value = 0.76, standard error = 0.1 1. Caps indicates capsular; E Caps, early capsular; TA, tissue approximation; SB, springy block; S, spasm; and E, empty.

ed the intensity of their pain from 0 to 10 following each ive physiologic motion end-feel test. On the scale, 0 meant no pain and 10 meant extremely severe pain. Only the first examination data obtained by each evaluator were used in this report, as would occur in the clinical setting, and all subjects were used in the analysis. Intra-rater and inter-rater reliability was examined. For the knee, intra-rater kappa coefficients (mean = 0.88, SD = 0.17, n = 23) varied from 0.76 to 1.00 with 83 to 100% agreement. An example of a contingency table for intra-rater ive knee flexion is included (Table 3). At the shoulder, intra-rater kappa coefficients (mean = 0.83, SD = 0.16, n = 28) varied from 0.65 to 0.92 with 86 to 96% agreement. Inter-rater reliability kappas (knee mean = 0.21, SD = 0.31, n = 17, and shoulder mean = 0.45, SD = 0.16, n = 18) were below 0.47 with between 35 and 89% agreement. Based on the literature on pain scales, we assumed that the pain reports from s u b jects were reliable.2.4p8

Data Analysis Data were analyzed using one-way analysis of variance (ANOVA) using end-feel category as the factor with 6 levels to determine whether the observed dif-

ferences at the knee or the shoulder among the pain means, based on end-feel category, were greater than expected by chance. For knee and shoulder motions in which significant differences were found, post hoc Tukey's Honestly Significant Difference (HSD) tests were used to determine exactly which groups differed. The alpha level was set at 0.05 for all analyses. For one motion, knee flexion, variances were not homogeneous. A Kruskal-Wallis (KW) one-way ANOVA was further used to analyze knee flexion. The ANOVA for shoulder glenohumeral abduction showed an omnibus difference and Tukey's HSD was not sensitive enough to find where there were differences; therefore, the Least Significant Difference (LSD) test was used to analyze (post-hoc) the shoulder glenohumeral abduction data set." To check whether reliability of the data affected the results, we analyzed data from all subjects, as well as from only those subjects on whom first and second evaluations produced the same end-feels. The analyses were performed using a Power Macintosh computer (Apple Computer Inc, Cupertino, Calif), SPSS 6.1.1 statistical software (SPSS Inc, Chicago, Ill) and a Claris Works 4.0 spreadsheet (Claris Corporation, Santa Clara, Calif).

RESULTS The relationship between normal and abnormalpathologic end-feels was the same at the knee and the shoulder. The magnitude of the mean pain was significantly higher for spasm end-feel than for c a p sular end-feel for knee extension (Table 4 and Figure 1). For knee flexion, early capsular, spasm, springy block, and empty end-feels were more painful than tissue approximation and capsular end-feels (Table 4 and Figure 1). Early capsular and empty end-feels were more painful than the capsular endfeel for shoulder full abduction and for shoulder glenohumeral abduction (Table 4 and Figure 2). Early capsular and spasm end-feels were more painful than the capsular end-feel for shoulder external rotation

TABLE 4. Tukey's HSD post hoc significant difference findings in mean (SD) pain comparing pathologic and normal end-feels.

Normal

Pathologic EC

S

SB

Knee Extension Flexion

3.1 7 (1.60) 4.8 (1.92)*

7.0 (0)' 6.5 (3.54)'

7.0 (0).

Shoulder Full Abd GH Abd ER IR HA

4.88 (2.55)' 2.5 (2.01)' 3.18 (2.71)' 4.62 (2.80)' 5.06 (2.67)'

3.0 (0) 9.0 (0)'

E 4.0 (0) 8.25 (0.35)' 6.33 (2.31)' 5.0 (0)' 4.0 (0) 4.0 (0) 4.0 (1.41)

TA

C

0.47 (1.06)

1.3 1 (1.93) 1.4 (1.99)

0.60 (1.34)

2.67 (2.38) 0.88 (1.63) 1.12 (1.94) 2.0 (2.19) 1.23 (2.25)

Abd indicates abduction; TA, tissue approximation; GH, glenohumeral; C, capsular; ER, external rotation; EC, early capsular; IR, internal rotation; S, spasm; HA, horizontal adduction; SB, springy block; and E, empty. Indicates a significant difference compared to a normal end-feel (P< .05). 516

J Orthop Sports Phys Ther.Volume SO-Number goSeptember 2000

Flexion

Extension

*

12

TA

Cap

E Cap

SB

I

Spasm Empty

Cap

E Cap Spasm Empty

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

FIGURE 1. Knee physiologic motion. Stacked mean pain ratings (black) and standard deviations (white) for each end-feel by motion at the knee. Starred knee physiologic motion pathologic end-feels are more painful than normal (capsular and tissue approximation) end-feels. End-feel categories: TA indicates tissue approximation; Cap, capsular; E Cap, early capsular; SB, springy block. k i n ratings for pathological end-feels are significantly greater than pain ratings for normal end feels.

from early capsular to spasm, springy block, and empty, as seen in Figure 1. Because of the small sample sizes in some categories, the power of some of the comparisons is low (Figure 3). In other cases, as with shoulder external rotation (Figure 2), a single individual with a high pain level (spasm end-feel) made the comparison significant (knee extension, one individual with a spasm end-feel; knee flexion, one individual with a springy block end-feel; shoulder glenohumeral abduction, one individual with an empty end-feel; and shoulder external rotation, one individual with a spasm endfeel) (Figures 1 and 2). Examining results using the data from the reliable subset (subjects with the same end-feel on both evaluations), the pattern of end-feel responses did not change, and the specific comparisons were almost

(Table 4 and Figure 2). The early capsular end-feel was more painful than the capsular end-feel for shoulder internal rotation, and the early capsular end-feel was more painful than tissue approximation and capsular end-feels for shoulder horizontal adduction (Table 4 and Figure 2). Tissue approximation and capsular end-feels were the only expected normal end-feels found at both the knee and the shoulder. No bone-to-bone end-feels were found at either t and would have been considered abnormalpathologic at these ts. Early capsular end-feels were more painful than capsular normal end-feels for both ts, suggesting pathology. Knee flexion was the sole motion in which all endfeel categories defined in the study were found (except bone-to-bone). The pain associated with the a b normal-pathologic end-feels increased in intensity

Full Abduction

Cknobumcrd Abduction

External Rotation

lntcrnal

Horizontal

Rotation

Adduction

FIGURE 2. Shoulder physiologic motion. Stacked mean pain ratings (black) and standard deviations (white) for each end-feel by motion at the shoulder. Starred shoulder physiologic motion pathologic end-feels are more painful than normal (capsularand tissue approximation)end-feels. End-feel categories: Cap indicates capsular; E Cap, early capsular; TA, tissue approximation. Pain ratings for pathological end-feels are significantly greater than pain ratings for normal end feels. J Orthop Sports Phys Ther *Volume SO. Number 9. September 2000

517

Physiologic Motions

n E l Knee Ext

W Knee Flex H Shld FA

H Shld CHA H S h l d ER Shld IR Shld H A

TA

Cap

ECap

Spasm

SB

Empty

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

FIGURE 3. Number of subjects in each end-feel category by t motion. End-feel categories: TA indicates tissue approximation; Cap, capsular; E Cap, early capsular; SB, springy block. Physiologic motions: Ext indicates extension; Flex, flexion; Shld, shoulder; FA, full abduction; GHA, glenohumeral abduction; ER, external rotation; IR, internal rotation; HA, horizontal adduction.

identical except for shoulder full abduction and glenohumeral abduction (Figures 4 and 5). The abnormal-pathologic end-feels were associated with more pain, but were not always painful, and normal end-feels were not always pain-free, as would be expected (Table 5). Across both ts, when a b normal-pathologic end-feels did not produce pain, the end-feels were all early capsular. When normal end-feels produced pain, all but 3 of the end-feels were capsular. The other 3 were tissue approximation end-feels with knee flexion. Both of these results can be explained by pain interpretation variability among individuals. All 3 end-feels (capsular, tissue approximation, and early capsular) can create a stretching or squeezing sensation in tissues that may be felt as painful by some individuals and nonpainful by others. In a clinical situation, practitioners would compare the findings with the opposite extremity. These comparisons were done in this study. In subjects who indicated pain but had a normal end-feel, we suspect that they interpreted the stretch of the capsule (capsular) or tissue compression (tissue a p proximation) as painful and that they would report similarly for the uninvolved side. Subjects with an early capsular end-feel without pain may not have interpreted the stretch as painful, but would be expected to have a loss in range of motion between the involved and uninvolved sides. In this study, 7 of the 11 subjects clearly had range of motion loss in the involved side compared with the uninvolved side, and in the other 4 subjects, the difference in range of motion was so small as to be within measurement error for range of motion. Other findings would add to whether pathology were present or not, because no examination finding should be interpreted in isolation. These findings our hypothesis that at the knee and the shoulder, subjects who had early capsular, spasm, springy block, and empty end-feels had

more pain during end-feel testing than subjects with normal end-feels.

DISCUSSION

Subjects The subjects recruited for this study were recruited mainly from the university's student community. Most fell within the age range of 20-30 years and had a median duration of pain of one year. The knee problems were equally distributed between the right and left sides and the shoulder problems were mostly right-sided. The results from this study may not relate to other age ranges or to a median duration of symptoms other than one year's time. More right shoulder involvement could be expected because right-hand dominance was more common and, therefore, more strongly associated with developing painful shoulder problems. These are areas for further research.

End-feels and Pain Our hypothesis that some abnormal-pathologic end-feels are significantly more painful than normal end-feels was ed for all motions at each t (Figures 1 and 2). There may also be a pattern of increasing intensity of pain responses within the endfeel categories. Early capsular abnormal-pathologic end-feels were more painful than normal capsular end-feels found with knee extension and all shoulder motions and normal tissue approximation end-feels found with knee flexion and shoulder horizontal adduction (Figures 1 and 2). Within the abnormalpathologic end-feels, when all categories were present except bone-tebone (knee flexion, Figure 1). pain intensity increased from early capsular to spasm, springy block, and empty end-feels. Within this orJ Orthop Sports P h y Ther-Volume 30. Number 9-September 2000

Flexion

Extension

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

FIGURE 4. Knee physiologic motion. Stacked mean pain ratings (black) and standard deviation (white) for each end-feel by motion at the knee in the reliable subset. Starred knee physiologic motion pathologic end-feels are more painful than normal (capsular and tissue approximation) end-feels in the reliable subset. End-feel categories: TA indicates tissue approximation; Cap, capsular; E Cap, early capsular; SB, springy block. Pain ratings for pathological end-feels are significantly greater than pain ratings for normal end feels.

dering of increased pain responses with the 4 abnormal-pathologic end-feels, we expected that an early capsular end-feel would be the least painful. An early capsular end-feel "when felt in conjunction with a capsular pattern of restriction and in the absence of indicates significant inflammation or effusion capsular fibro~is."'~(*~) As described, stretch should be felt with this end-feel and some people might consider the stretch painful because of the abnormal nature of the tissue being stretched. In the individual motion analyses, there were no statistical differences in pain intensity among the a b normal-pathologic end-feels, mostly due to small sample size. To increase the sample sizes in the end-feel categories and to explore potential differences in pain intensity among the abnormal-pathologic end-

.. .

Full Abduction

Glenohumeral Abduction

feels, we pooled the first measure of all the end-feels for both the intra-rater and inter-rater conditions for both motions at the knee and the 5 motions at the shoulder for all subjects (total of 310 end-feels). We believe this aggregation is justified because abnormal-pathologic end-feels are the same at each t. We found significant differences among end-feels F, = 26.34. Tukey's HSD showed that all abnormal-pathologic end-feels were significantly more painful than normal end-feels and that spasm and empty end-feels were significantly more painful than an early capsular end-feel. This result reinforces the observation that an early capsular end-feel is less painful than the other abnormal-pathologic end-feels. Ordering the intensity of the 3 remaining abnormal-pathologic end-feels, spasm, springy block, and External Rotation

Internal Rotation

Horizontal Adduction

FIGURE 5. Shoulder physiologic motion. Stacked mean pain ratings (black) and standard deviations (white)for each end-feel by motion at the shoulder in the reliable subset. Starred shoulder physiologic motion pathologic end-feels are more painful than normal (capsular and tissue approximation)endfeels in the reliable subset. End-feel categories: TA indicates tissue approximation; Cap, capsular; E Cap, early capsular; SB, springy block. Pain ratings for pathological end-feels are significantly greater than pain ratings for normal end feels. J Orthop Sports Phys Ther.Volume 3O.Number 9.September 2000

TABLE 5. Number of subjects in combined normal versus combined pathologic end-feel categories by t motion and if pain was present or absent. Physiologic motion Knee

Shoulder

-- -

End-feel and pain

EXT

FLEX

FA

GHA

ER

IR

HA

Normal EF, no pain Normal EF with pain Pathologic EF, no pain Pathologic EF with pain

18 14 0 8

21 9 0 10

9 21 1 15

24 10 3 9

17 8 5 16

10 14 2 20

25 10 0 11

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

EXT indicates extension; FLEX, flexion; FA, full abduction; CHA, glenohumeral abduction; ER, external rotation; IR, internal rotation; HA, horizontal adduction; and EF, end-feel.

empty is speculative. All 3 end-feels are suggestive of painful pathologies, but to indicate specifically which pathology would be the most painful would be dependent on pain behaviors that vary in individuals. A spasm end-feel can occur with capsular restrictions indicating "some degree of synovial inflammation" with "pain felt at the point of restriction."'O(~A springy block end-feel suggests internal derangement "such as displacements of pieces of torn menisci and ~artilage"~(@~) or bone loose bodies. Cyriax states "the pain is localized""~~~) with springy block endfeels. Empty end-feels typically are rare. The end-feel is present with acute bursitis at the shoulder and "other painful extraarticular lesions such as neoplasms."'O(fl) Additional research may reveal a specific ordering of pain within the abnormal-pathologic end-feel categories. We think that our finding that an early capsular end-feel was the least painful of the abnormal-pathologic end-feels does reflect the pathophysiology that is present when an early capsular end-feel is found. The results of this study were expected because knee or shoulder pathology suggests changes within the normal physiology of one or more tissues (inert or ile) surrounding the t complex. When changes in physiology occur, pain is frequently a syrnpt~m.'~ Subjects were chosen for the study because they had symptoms related to knee or shoulder problems suggestive of pathology. In spite of our efforts to include subjects with a wide variety of problems, some end-feels were rare, leading to low numbers of observations for some comparisons. Coupled with comparisons in which a single subject had extreme pain, these results must be interpreted with caution. We believe, however, that the pattern of results is robust. In general, abnormal-pathologic end-feels produce more pain than normal end-feels.

Reliability We believe that reliability of the end-feel data did not affect the results because the pattern of end-feel responses did not change, and the specific comparisons were almost identical when the reliable subset was used. The only differences were for shoulder full

abduction (the order was the same but none were significant) and glenohumeral abduction (only the empty end-feel category was significant) (Figures 4 and 5). The decrease in the number of significant comparisons was likely due to the decrease in the sample sizes in the reliable subset. More research is needed to confirm these findings with the same and other ts. In examining the inter-rater reliability, most errors were found between the categories of capsular and early capsular end-feels. Whether these end-feels produce pain may depend on the vigor of the overpressure applied. Care must be taken to differentiate stretch from pain and to attend to even slight limitations in motion.

End-feel Names There is no evidence to suggest that the end-feel names are valid representations of the primary tissue that limits movement at a t. Caution must be used, therefore, in interpreting end-feel names as indicators of tissues that are restricting motion at ts. Like Cyriax, Paris specifically categorizes endfeels based on the anatomic structure limiting the range of m o t i ~ n . ~Kaltenborn, .'~ on the other hand, suggests using descriptive labels for end-feels including soft, firm, or hard but, again, dependent on the anatomy of the t.g Refinement of definitions is also an area for further research. Clinicians must consider that musculoskeletal involvement may be present if the end-feels defined as abnormal-pathologic by Cyriax are found during ive physiologic movement testing. Early capsular end-feels may be suggestive of ligamentous or capsular fibrosis, but the nature of the relationship between pain and mobility limitations is unclear. Further research is also needed to determine if similar end-feel findings occur with ive physiologic motion at other ts.

CONCLUSION These data show that abnormal-pathologic endfeels can be associated with more pain than normal end-feels during ive physiologic motion testing at J Orthop Sports Phys Ther *Volume SO. Number 9. September 2000

the knee and shoulder. A trend toward an increasing intensity o f pain responses within the end-feel categories was found. These results provide for the concept o f abnormal-pathologic versus normal end-feels in evaluation o f patients with musculoskeletal problems. When abnormal-pathologic end-feels are present, dysfunction should be suspected within the tissues surrounding that j o i n t complex. End-feel testing, part o f ive movement assessment, is a n important part o f the physical examination, providing information that may indicate the presence o f tis sue pathology. Further research is necessary t o continue t o examine end-feels and their relationship t o pathology.

6.

7. 8. 9. 10.

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

ACKNOWLEDGMENTS We thank o u r research assistant, evaluators, and the subjects who participated in this study. We also thank Carol Tercyak, PT,MS, for reviewing the manuscript, and Ellen Humphrey, PT,MA, OCS, Yolanda Aldrete, and Phyllis Koerner f o r help with manuscript preparation.

REFERENCES 1. Bijl D, Dekker J, van Baar ME, et al. Validity of Cyriax's concept of capsular pattern for the diagnosis of ostoearthritis of hip andlor knee. Scand J Rheumatol. 1998;27; 347-351. 2. Boeckstyns MEH, Backer M. Reliability and validity of the evaluation of pain in patients with total knee replacement. Pain. 1989;38;29-33. 3. Cyriax J. Textbook of Orthopaedic Medicine Volume I. Diagnosis of Soh Tissue Lesions. 8th ed. London: Bailliere Tindall; 1982. 4. Downie WW, Leatham PA, Rhind VM, Wright V, Branco JA, Anderson JA. Studies with pain rating scales. Ann of Rheum Dis. 1978;37:378-381. 5. Franklin ME, Conner-Kerr T, Chamness M, Chenier TC,

J Orthop Sports Phys Ther .Volume SO. Number 9 September 2000

11.

12.

13. 14. 15.

16. 17.

Kelly RR, Hodge T. Assessment of exerciseinducedminor muscle lesions: the accuracy of Cyriax's diagnosis by selective tension paradigm. ) Orthop Sports Phys Ther. 1996; 24:122-129. Fritz JM, Delitto A, Erhard RE, Roman M. An examination of the selective tissue tension scheme, with evidence for the concept of a capsular pattern of the knee. Phys Ther. 1998;78:1046-1061. Hayes KW, Petersen C, Falconer J.An examination of Cyriax's ive motion tests with patients having osteoarthritis of the knee. Phys Ther. 1994;74:697-707. Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: a comparison of six methods. Pain. 1986;27:117-126. Kaltenborn FM. Mobilization of the Extremity )oink: Examination and Basic Treatment Techniques. 3rd ed. Oslo: Olaf Norlis Bokhandel; 1980. Kessler RM, Hertling D. Management of Common Musculoskeletal Disorders: Physical Therapy Principles and Methods. Philadelphia: Harper & Row; 1983:44-48, 99101. Ombregt L, Bisschop P, ter Veer HJ, Van de Velde T. Clinical examination of the shoulder. In: Ombregt L, Bisschop P, ter Veer HI, Van de Velde T. A System of Orthopaedic Medicine. London: WB Saunders Company Ltd; 1995: 215-223. Patla CE, Paris SV. Reliability of interpretation of the Paris classification of normal end feel for elbow flexion and extension. lournal of Manual and Manipulative Therapy. 1993;1:60-66. Paris SV, Patla C. E l course notes: Extremity dysfunction & manipulation. Atlanta, Ga: Patris Inc; 1988:86-88. Pellechia GL, Paolina J, Connell J. lntertester reliability of the Cyriax evaluation in assessing patients with shoulder pain. J Orthop Sports Phys Ther. 1996;23:34-38. Petersen C. Active and ive Movement Testing of the Extremities, Spine, Pelvis and Temporomandibular t. Chicago, Ill: Programs in Physical Therapy; 1999:7lO,l3,87-88. Petty NJ, Moore AP. Neuromuscular Examination and Assessment: A Handbook for Therapists. Edinburgh: Churchill Livingstone; 1998: 6, 8. Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. Connecticut: Appleton & Lange; 1993:400-402,429-432.

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Invited Commentary Medicine has long employed palpation and manual exploration of skeletal and visceral structures in order to determine the presence or absence of pathology or dysfunction. For example, in individuals with suspected appendicitis, McBurney's point is palpated and the pain response is noted. Despite the continual technological explosion in medicine, the patient history and physical examination are undergoing a renaissance of sorts and are again being emphasized in clinical diagnosis. The study by Petersen and Hayes examined Cyriax's system of selective tension,' which is a manual evaluation method commonly employed by physical therapists. The authors have attempted to provide construct validity to this concept when applied to t end-feels in subjects with pain at the knee and shoulder. The authors hypothesized that abnormal-pathologic end-feels would occur in a pathologically involved t, and the pain associated with an abnormal-pathologic end-feel would be greater than the pain found with a normal endfeel. The authors reported that the 5 shoulder positions and knee flexion resulted in increased pain when an early capsular pattern was present. The authors also attempted to draw inferences from other categorical comparisons, but these may be invalid. In particular, a major limitation in this study was the extremely low (1-3) number of cases in 3 of 6 end-feel categories (spasm, springy block, and empty end-feel). In addition, the low number in several of the end-feel categories had a significant effect on statistical power for between group comparisons. At the outset, there was no mention of what a clinically meaningful difference in pain between end-feel categories should be, and the number of subjects per category needed to detect this difference, if it exists. Furthermore, the authors state "because of the small sample sizes in some categories, the power of some of the comparisons is lo~."~(p"')However, the authors never provide the reader with a value for that "low" statistical power. More concerning is the multiple between category comparisons. For instance, the authors state "In other cases, as with shoulder external rotation, a single individual with a high pain level (spasm end-feel) made the comparison significant. . . "2(p"7) In fact, they describe 5 of these cases in which there is only one individual in the category; however, the comparisons were significant. Some of these findings are likely spurious,

particularly in light of the fact that multiple comparisons (21 in Table 4) were made with no attempt to adjust the alpha level in order to prevent a Type I error; some of these findings are just as likely due to chance."t is questionable whether or not post-hoc tests performed with only one subject and, therefore, with no variance is a reasonable a p proach. Any discussion of the spasm, springy block, and empty end-feel, therefore, should be viewed with extreme caution. A potentially more meaningful analysis would be to know whether or not the subjects with end-feel categories with low prevalence (spasm, springy block, and empty end-feel) had diagnoses that differed from those subjects with early capsular patterns and if their eventual treatment outcomes were grossly different. This would provide for the concept that even though the prevalence of these categories is low, the classification of patients into these categories either assists with identifying "red flag" diagnoses (ie, tumor, infection, etc) or results in a distinctly different treatment approach. If not, then do we need all these categories, or is normal and abnormal enough? The purpose of a diagnostic or classification system is to identify subgroups of patients who are more likely to respond to a particular intervention. The selective tension scheme is designed to systematically stress various t-related structures. The interpretation of the examination then theoretically guides the therapist's treatment decisions. However, the determination of the "end-feel" by the therapist is a multifactor decision that s for both the patient's response and the therapist's interpretation of the end-range feel; in clinical practice, this is likely influenced by knowledge of the patient's present symp toms and other historical factors. In this study, the therapists recorded the end-feel and the subjects reported the amount of pain they felt. The therapist had no knowledge of the patient's history or present symptoms. This was purposeful in order to minimize bias from the history, which may influence the therapist's end-feel decision. It does limit the generalizability of the results. The process of appraising the elements of a patient's history and physical examination, both part of the orthopaedic physical therapy practice, is critical to our professional growth. Research can provide evidence that either validates or refutes examination systems; however, it is difficult to interpret if the present paper provides proof or disproof for the seJ Orthop Sports Phys TherrnVolume SO. Number 9. September 2000

lective tension concept. At best, I believe we are left with the idea that the selective tension concepts may have some merit. Timothy W. Flynn, PT, PhD, OCS, FAAOMPT Associate Professor and Research Director U.S. Army-Baylor Graduate Program in Physical Therapy 3151 Scott Road Fort Sam Houston, TX 78234-6138

REFERENCES 1. Cyriax J. Textbook of Orthopaedic Medicine, Volume 1: Diagnosis of Soft k u e Lesions. 8th ed. London, England: Bailleire Tindall; 1982. 2. Petersen CM, Hayes KW. Construct validity of Cyriax's selective tension examination: association of end-feels with pain at the knee and shoulder. ] Orthop Sports Phys Ther. 2OOO;3O:Sl2-521. 3. Portney LG, Watkins MP. Foundations of Clinical Research. Applications to Practice. Norwalk, Conn: Appleton & Lange; 1993.

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Invited Commentary The paper "Construct Validity of Cyriax's Selective being applied to test inert structures and resisted tests to stress contractile structures. Tension Examination: Association of End-feels With The application of ive movements reveals a Pain at the Knee and Shoulder" by Cheryl Petersen, specific sensation through the operator's hands when PT, MS, and Karen Hayes, PT, PhD, examines the the end of available range has been reached. Dr Cyconcept of end-feel and relates the normal to the abriax termed this sensation the "end-feel." Petersen normal-pathogenic using pain experienced by suband Hayes use classifications of end-feels according jects as an indication of pathology. The knee and to Cyriax, Paris, and Kaltenborn, but the Cyriax classhoulder are investigated within this preliminary study. The methodology and discussion consider oth- sification, and the subsequent terminology adopted within the paper, have been developed in the curer interpretations of end-feel, but are rooted in the rent teaching of orthopaedic medicine and may be classification expounded upon by Dr James Cyriax, confusing to those who have studied the specialism FR, which is notably relevant to the assessment procedures used in orthopaedic medicine. Cheryl Pe- more recently. As mentioned in the paper, Cyriax classified normal end-feels as "bone-to-bone," "tissue tersen and Karen Hayes' attempt to substantiate Dr approximation," and "~apsular,"~ and the authors' Cyriax's theories should be applauded. explanations are an accurate representation of when Dr Cyriax's claim that "all pain arises from a lethey exist. Current orthopaedic medicine teaching, sion, all treatment must reach the lesion, and all however, tends to describe normal end-feels as hard treatment must exert a beneficial effect on the le(bone-to-bone), soft (tissue approximation), and elassionW4(pJ) has been handed down through his teachtic (capsular), describing the elastic resistance proing years, as has his assertion that it was for his asduced in the inert tissues at the end of range,' and sessment procedures that he wanted to be known which, in the norm, could include tendon or ligaand ed. His pragmatic and logical a p ment rather then confining itself to capsular resisproach towards establishing the "source" of the tance only. symptoms led him to state that normal tissue should In the abnormal end-feels that Cyriax referred to function painlessly, whilst traumatised or inflamed tissue would It followed then, that if appropriate as "capsular," "spasm," "springy," and "empty,""t is the abnormal "capsular" end-feel that I believe tension could be applied to specific tissues, then the outcome, in of pain production, would be able needs further discussion. Cyriax noted that each individual t develops a pattern of pain and limitation to incriminate or eliminate the structure from the within its movements when the capsule is irritated. clinical "inquiry." The study by Petersen and Hayes s the theories underpinning the use of selec- This pattern was demonstrated by limitation of movemen& in a fixed proportion, which he described as tive tension as a diagnostic tool, but some discussion the capsular pattern, and he propounded that each is required to clarify the clinical significance of its t has its own specific pattern. The significance of conclusions. the capsular pattern is that it denotes that an "arthriDr Cyriax classified the soft tissues into inert and tis" is present, using the term as a generic for any contractile structures; inert implying t capsules, cause of inflammation of the t capsule, be it deligaments, fascia, bursae, etc, and contractile as musgenerative, traumatic, or inflammatory disease. cle, tendon, and attachments to bone.5 The term seAt the knee, the capsular pattern is represented by bctive tension denoted that appropriate tension should be applied to tissues, with ive movements proportionally more limitation of flexion than of exJ Orthop Sports Phys Ther -Volume 30. Number 9. September 2000

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

tension. The capsular pattern of the shoulder is described as proportionally more limitation of glenohumeral external rotation, less of abduction, and least of medial rotation. The concept of the capsular pattern is important in context of the study by Petersen and Hayes and helps to explain several of the observations. Revisiting the notion that an end-feel is a specific sensation imparted through the operator's hands, the abnormal-pathologic capsular end-feel will tend to feel "harder" when compared with the asymptomatic side, even at an early stage in the pathological process. This is attributed to the intervention of involuntary muscle spasm that protects the capsule from being stretched, long before capsular contracture has had a chance to d e v e l ~ pCapsular .~ contracture and fibrosis provide the next stage in contributing to the abnormal end-feel? and the "hard" endfeel becomes more apparent, again compounded by protective muscle spasm. I doubt though, that within the age group studied by Petersen and Hayes, true capsular changes have developed. Only in the most advanced stages of osteoarthrosis would degenerative changes within the articular cartilage or bone with osteophyte formation be likely to present a true bone-to-bone end-feeL3 The point that should be emphasised is that it is essential to compare with the asymptomatic side in order to fully exploit the relevance of the test and its clinical usefulness. Cyriax linked the relationship of pain and capsular limitation as a guide for treatment techniques that involved the use of ive stretching to increase range and to reduce pain."F) The presence of pain before capsular resistance is experienced indicates that an active or irritable lesion is present that renders the t unsuitable for stretching. Hyperirritability may be demonstrated by the presence of the "empty"' endfeel referred to earlier, but I would suggest that this may demand further investigation. Pain synchronous with limitation and with some preserved elasticity in the end-feel indicates that the t is suitable for stretching. However, in the presence of the "hard" end-feel of muscle protection, the advice would be to employ other treatment modalities before introducing stronger stretching mobilisations. Capsular limitation before the onset of pain acts as a guide that the t will tolerate strong ive stretches, particularly effective in treating the less irritable, lower grade "arthritis" in the shoulder and hip. As mentioned above, even in the earliest stage of capsulitis, a "hard" capsular end-feel may exist due to the involuntary muscle activity that acts to prevent the inflamed tissues being stretched as a response to the pain felt on specific ive movements. In p r e p aration for our own text,7 our literature search revealed the theory of Eyring and Murray,%hich stated that the development of the capsular pattern could be associated with t effusion, causing the

t to assume an antalgic position of ease, with movements out of this position causing pain and protective spasm. Eyring and MurrayG noted a possible relationship between intra-articular pressure and pain, and conducted a series of experiments to determine the position of minimum pressure in various ts. It was observed that symptomaticts with effusion spontaneously assume a position of minimum pressure, which coincides with that of minimum pain. This provides a possible explanation for the specific capsular patterns and s for the development of a "hard" end-feel associated with pain on specific movements, even in the early capsul i t i ~In . ~ protecting the painful, inflamed t, the painful movements are comparatively underused and the limitation in the capsular pattern is further compounded by the ensuing capsular contracture and fibrosis, as attributed by Petersen and Hayes to Cyriax,3 and as ed by B ~ n k e r . ~ Petersen and Hayes note the affective aspects of pain experienced and observe that this could explain why subjects with an early capsular pattern at the shoulder do not complain of pain at the end of range of motion on external rotation at the shoulder, whilst subjects with apparent asymptomatic shoulders may resent the capsular stretch and complain of pain at end of normal range of motion. Petersen and Hayes propose that the evaluators' skills in applying overpressure might be a factor in this, but their findings broadly the intrarater and interrater reliability of the study for the shoulder. Cyriax2 noted that "pain is not always an essential component of limited movement in arthritis." Using the hip as an example, he observed that radiologically confirmed arthritis, with considerable restriction of range of motion in the capsular pattern, may coexist with complete absence of pain when the capsule is stretched quite hard.3(p55)This may also act as an example of the resistance before the experience of pain (as referred to earlier). The comparison with the contralateral side would satisfactorily provide a check on the clinical significance of these findings, but it is not clear from Petersen and Hayes that this was always performed. I question the specific movements selected at each t studied as being representative. At the shoulder, external rotation, internal rotation, and the end of ive elevation through flexion would conceivably stress the capsule to provide a capsular end-feel. Socalled glenohumeral abduction is used more as a measure of range of movement at the glenohumeral t, observing the commencement of scapular movement, and does not provide a true end-feel.' Horizontal adduction is taught by the Society of Orthopaedic Medicine as an accessory test for impingement of, for example, the upper fibres of subscapularis or the subdeltoid bursa, or compression of the acromioclaviculart. I accept, however, that pain J Orthop Sports Phys Ther .Volume SO. Number 9. September 2000

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

may produce a protective muscle spasm, which may produce an abnormal "hard" capsular end-feel. On applying ive extension to the knee, an initial overpressure is usually applied to test for range of movement, whilst a small crisp drop or "bounce home" into extension applied to a normal knee should produce a harder, normal bone-to-bone endfeel,' similar to that encountered as the normal endfeel of ive elbow extension. The concept of the "capsular" pattern is a valuable tool in clinical diagnosis and the end-feel provides an effective guide for treatment selection. I note the suggestion of Sims: based on the findings of Bijl et all with reference to the hip, that little evidence was found to the classically described capsular pattern as presented by Cyria~.~5 However, my personal clinical experience of 26 years is thoroughly ive of the classical capsular pattern at the hip, the knee, and shoulder specified in the study by Petersen and Hayes, and as given for other for diagnosis and treatment. A balance should be found between finding evidence to our clinical efficacy and establishing evidence to refute its benefit, and I fear that the emphasis is more often on the latter. Petersen and Hayes' demonstration that abnormal-pathologic endfeels produce more pain than normal end-feels has provided a valid endorsement for the link between abnormal-pathologic end-feels and regional dysfunction. They acknowledge that further research is needed to examine end-feels and their relationship to pathology, and, in context of the complete Cyriax assessment, the specific tissue affected. Petersen and Hayes' introduction explains that "part o f ' Cyriax's selective tension theory involves the concept of end-feel testing during ive movements. The study, however, is ultimately focusing on capsular considerations where the end-feel is particularly important in diagnosis and treatment. Later, Petersen and Hayes state that "other findings" would add to whether pathology were present or not, because no examination finding should be taken in isolation. Other inert tissues such as ligament or bursa, and even tendonitis, may produce pain on ive testing, but do not usually present with an abnormal-pathologic end-feel. Points from the patient's history and analysis of other tests lead to confirmation of clinical diagnosis. In relation to the shoulder, Pellecchia et al,R in looking at the in-

J Onhop Sports Phys Ther .Volume SO. Number 9 September 2000

tertester reliability of the Cyriax evaluation as a whole, found it to be highly reliable in the assessment of patients with shoulder pain, facilitating the identification of diagnostic categories. I urge those therapists who may have come to the approach through the original work of Dr Cyriax to reacquaint themselves with its tenets in light of the developments that are continuing in orthopaedic medicine and the continued expansion of its evidence base through studies such as this. The paucity of quality research to provide for much of our work as physical therapists is a frustration to all in our profession. The Society of Orthopaedic Medicine has funds available that are intended to projects that set out to increase the evidence base of musculoskeletal therapy, and welcomes applications. On behalf of the Society, I am pleased to have had the opportunity to provide a commentary on this paper, and I thank Dr Di Fabio for his invitation. Elaine Atkins, MA, Grad Dip Phys, MCSP SRP Course Principal, Society of Orthopaedic Medicine 154 High Road Woodford Green Essex, England IG8 9EF Society of Orthopaedic Medicine www.soc~rthemed.org

REFERENCES 1. Bijl et al. Validity of Cyriax' concept of capsular pattern for the diagnosis of osteoarthritis of hip andlor knee. Scand) Rheumatol. 1998;27:347-35 1. 2. Bunker TD. Frozen shoulder: unravelling the enigma. Ann R Coll Surg Engl. 1997;79:21 G213. 3. Cyriax J. Textbook of Orthopaedic Medicine Volume 1. Diagnosis of Soft Tissue Lesions. 8th ed. London: Bailliere Tindall; 1982. 4. Cyriax J. Textbook of Orthopaedic Medicine. Volume 2. Treatment by Manipulation, Massage and Injection. 11th ed. London: Bailliere Tindall; 1984. 5. Cyriax JH, Cyriax PJ. Illustrated Manual o f Orthopaedic Medicine. Oxford: Butterworths; 1983. 6. Eyring EJ, Murray WR. The effect of t position on the pressure of intra-articular effusion. ) Boneloint Surg. 1964; 46-A:l235-1241. 7. Kesson M, Atkins E. Orthopaedic Medicine--A Practical Approach. Oxford: Butterworth Heinemann; 1998. 8. Pellecchia GL, Paolina J, Connell J. lntertester reliability of the Cyriax evaluation in assessing patients with shoulder pain. ) Orthop Sports Phys Ther. 1996;23:43-48. 9. Sims K. Assessment and treatment of hip osteoarthritis. Manual Therapy. 1999;4:136-144.

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Author Response We thank Ms Atkins and Dr Elynn for a dialogue According to Cyriax, end-feels are based on the that provides incentive to continue researching selec- anatomy of the t ~ o m p l e x * ( p pand, ~ ~ ~contrary ) tive tension testing and other physical therapy exami- to Ms Atkin's suggestion that knee and elbow extennation and treatment interventions. We note that sion end-feels are the same, the end-feel for knee exboth commentators appear to believe that research is tension may be slightly different than elbow extenperformed to "prove," "substantiate," or justify our sion due to the presence of the meniscii at the knee. practice. To the contrary, the purpose of research is The diierent definitions of end-feels that Ms Atkins a quest for truth, and all outcomes, both positive documents reflect current orthopaedic medicine teachand negative, can serve to inform decision making. ing that varies from Cyriax's original work. Based on The results of individual research studies may s u p research and clinical observation, change is construort or refute specific hypotheses and contribute to tive and inevitable, but these altered definitions cannot research and a larger body of knowledge, but no sin- be attributed to Cyriax. Ms Atkins stated that there are gle study can be used as proof that the intervention still 3 normal end-feels, but she uses diierent nomenor assessment is effective. clature: hard (bone-to-bone), soft (tissue approximaWe believe that evidence from a systematic collection), and elastic (capsular). Three abnormal-pathologtion of data is more meaningful than clinical obseric end-feels remain unchanged from Cyriax's definivation or experience. Elaine Atkins, for example, tions of spasm, springy block, and empty end-feels. Ms seems to believe that our study "substantiates" CyAtkins' description and explanations of her fourth a b riax's approach to selective tension testing, but disnormal-pathologic end-feel, "hard," seem to be based misses Bijl's study,' which questions the capsular paton Cyriax's concept of the capsular pattern and endtern at the hip, by stating that her clinical observafeels found during the stages of arthritis, but they diftions are superior to their data. Clinical observation fer from those of Cyriax. Ms Atkins does not use the is a good source of testable hypotheses, but research "spasm" and "capsular" as abnormalpathologic is needed to make datadriven decisions regarding end-feels with arthritis. She uses "hard" end-feel to dethe best examination techniques and treatment inter- scribe the muscle spasm that occurs during acute or ventions used to treat orthopaedic problems. subacute arthritis, the capsular changes that occur with Cyriax was logical and very insightful, but we must progression of arthritis and the end stage changes of continue to examine his and other pioneer work in or- articular cartilage degeneration and osteophyte formathopaedic medicine. Ms Atkins demonstrates an evolution. Our concern with Ms Atkins' use of only one type tionary process of changing an existing system. When of end-feel (hard) to describe the entire progression of clinicians change the principles or interpretations of a pathologic arthritic changes is that it fails to guide insystem such as that promoted by Cyriax, as Ms Atkins tervention. We believe that a spasm end-feel with acute does, it suggests that their experience with the system or subacute problems must be treated differently than causes them to question some of the premises. This a hard end-feel (Cyriax's capsular end-feel, early in the process is entirely appropriate, but those changes range of motion), and, in particular, a hard bone-toshould not be attributed to the originator of the sys bone end-feel. The original Cyriax abnormal-pathologic tem. Because so many clinicians cite Cyriax for their end-feel categories provided important and necessary practices, we decided to return to the basic tenets of information relative to intervention. Cyriax's work, without changing his examination Ms Atkins also expressed concern that the nature of scheme. the end-feel cannot be determined without comparison Ms Atkins questioned the movements that we testwith the contralateral side. In our study, end-feels were ed at the knee and the shoulder, the definitions of determined in relation to the contralateral side, which the end-feels, and whether a true end-feel was asis reflected by the inclusion criterion of unilateral sessed for glenohumeral abduction. We used the mo- symptoms and is directly stated in our manuscript (first tions suggested by Cyriax for the knee and shoulder, paragraph of the Methods section, Subjects). Normality and we used his 6 categories of end-feels. The endfor each subject was referenced to the uninvolved side feel testing that was performed for the 5 shoulder during all examination procedures. and 2 knee physiologic motions was based on his Ms Atkins focuses much of her commentary on testing scheme. End-feel testing was completed for the capsular pattern and pain resistance sequence as glenohumeral abduction using the principles of stadefined by Cyriax. The intent of our study was to exbilization and overpressure. plore Cyriax's concept of end-feels by examining J Orthop Sports Phrj Ther.Volume 30. Number 9. September 2000

Journal of Orthopaedic & Sports Physical Therapy® ed from www.jospt.org at on November 20, 2013. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

whether or not self-reported pain was greater for ab- very small probability values for the omnibus ANOVAs. normal-pathologic versus normal end-feels. The study Although we chose to report that the probabilities of the significant omnibus analyses were < .05, the actual was not intended to investigate Cyriax's concept of a probabilities were all substantially smaller, ranging from capsular pattern or to investigate pain coincident .O1 to < .0001. To address Dr Flynn's concern, we with resistance as a measure of chronicity or related reanalyzed the data using Bonferroni post hoc multiple to interventions. comparisons, which adjust the alpha level for the numWe disagree with Ms Atkins' assertion that inert tisber of comparison^.^ In all analyses, except the analysis sues other than the capsule and contractile structures for glenohumeral abduction, the results were exactly may produce pain without an abnormal end-feel. She offers no data to this assertion, and in making the same. There were no differences among end-feels for glenohumeral abduction using an adjusted alpha it, she disagrees with Cyriax and other orthopaedic practitioners. For example, Cyriax indicates that muscle level for the post hoc tests. spasm will occur with a gastrocnemius b r e a ~ h . ~ ( p ~ ~ ) We Dr Flynn's suggestions regarding an With acute subdeltoid bursitis, Cyriax states that muscle analysis of the relationship between low prevalence abnormal-pathologic end-feel categories and their respasm is not felt but that patients have an empty endlated diagnoses, and an analysis of whether intervenfeel.'(^'^^) The elicitation of muscle spasm is documenttion outcomes would differ among the various diaged by Petty and Moore7(p4*) "as a result of nerve irritanoses. These questions were not part of the original tion or secondary to injury of underlying structures, purpose of our study and, consequently, the diagnosuch as bone, t or muscle." Kessler and Hersis and intervention outcome data are either incomtling9(pp41wm.4*7) indicate that practitioners should note plete or were not collected. These questions would the presence of protective muscle guarding, suggestive be the basis for excellent future research projects. of a spasm end-feel, during t play movement testDr Flynn is concerned about the omission of the ining at the knee, including varus-valgus and internalexternal rotation tests, which are considered ligamentous terview as part of the patient examination. We acknowledge the importance of the patient interview, which, in stress tests. Meadows5(W)reports a hard capsular endfeel (what we called early capsular end-feel) with "peri- most circumstances, generates the hypotheses related to the probable pathologies that are further tested durcapsular tissue hypomobility caused by arthrosis, adheing the patient examination. Those hypotheses, howevsions, or scarring" suggestive of ligamentous involvement, and an early spasm end-feel "caused by arthritis, er, create expectations of the results of portions of the patient examination. Our omission of the interview was grade 2 muscle tear, fracture near a muscle insertion, intentional to allow assessment of the end-feel without dural sleeve, or other meningeal compression and/or that expectation bias. inflammation." While most of these statements are We thank the commentators, and we urge all orbased on observation and require systematic data colthopaedic practitioners to take every opportunity to lection to test their validity, Ms Atkins appears to stand collect data about their practices to inform future outside the mainstream on this issue. decision making. Regarding our statistical analysis, we appreciate Dr Flynn's comments. He is correct in urging caution Cheryl M. Petersen, PT, MS when interpreting the comparisons among groups Karen W. Hayes, PT, PhD with small sample sizes. We indicated that such comparisons had low statistical power, but did not report the actual power. Computation of power analyses REFERENCES with sample sizes less than 10 can be ina~curate.~ To Bijl D, Dekker J, van Baar ME, et al. Validity of Cyriax's exemplify the low power, we provide, as an example, concept of capsular pattern for the diagnosis of hip andlor knee. Scand) Rheumatol. 1998;27:347-351. the comparison between the early capsular ( n = 3) J. Textbook of Orthopaedic Medicine Volume I. DiCyriax and empty ( n = 2) end-feels for shoulder horizontal agnosis of Soft Tissue Lesions. 8th ed. London: Bailliere adduction. Even with an effect size of 2.5, usually Endall; 1982. considered to be quite large, the comparison was Kessler RM, Hertling D. Management of Common Musnonsignificant at an alpha level of .05. The power of culoskeletal Disorders: Physical Therapy Principles and Methods. Philadelphia, Pa: Harper & Rowe; 1983. Effect sizes cannot this analysis was less than Kraemer HC, Thiemann S. How Many Subjects? Statistical be computed with single subjects because of the lack Power Analysis in Research. Newbury Park, Calif: Sage of variance, and results representing single subjects Publications; 1987. can certainly be spurious. We concur with Dr Flynn Meadows TS. Orthopedic Differential Diagnosis in Physical and reiterate our own caution that these results must Therapy: A Case Study Approach. New York, NY: McGrawHill; 1999. be interpreted with caution. Much larger samples are Norusis MJ. SPSS 6.7 Base System 's Guide, Pdrt 2. needed to corroborate our results. Chicago, Ill: SPSS, Inc; 1994. Regarding the multiple comparisons without an adPetty NJ, Moore AP. Neuromusculoskeletal Examination justment in the alpha level, we thought that this proceand Assessment: A Handbook for Therapists. Edinburgh: Churchill Livingstone; 1998. dure would not be necessary in the presence of the J Orthop Sports Phys Ther.Volurne 30.Nurnber 9.Septernber 2000