The Supine Hip Extensor Manual Muscle Test:
A Reliability and Validity Study
Jacquelin Perry, MD, Walter B. Weiss, MPT, Judith M. Burnfield, PhD, PT, JoAnne K. Gronley, DPT ABSTRACT. Perry J, Weiss WB, Burnfield JM, Gronley
JK. The supine hip extensor manual muscle test: a reliability and validity study. Arch Phys Med Rehabil 2004;85:1345-50. Objectives: To define the relative hip extensor muscle strengths values identified by the 4 grades obtained with a supine manual muscle test (MMT) and to compare these values with those indicated by the traditional prone test.
Design: Comparison of 4 manual supine strength grades with isometric hip extension joint torque; statistic–deter-mined interrater reliability, and analyses of variance identified between grade differences in torque.
Setting: Pathokinesiology laboratory.
Participants: Adult volunteers recruited from local commu-nity and outpatient clinics. Reliability testing: 16 adults with postpolio (31 limbs). Validity testing (2 groups): 18 subjects without pathology (18 limbs), and 26 people with clinical signs of hip extensor weakness (51 limbs).
Interventions: Not applicable.
Main Outcome Measures: Supine hip extensor manual muscle grade and isometric hip extension torque.
Results: Reliability testing showed excellent agreement (82%). Subjects with pathology had significant differences in mean torque (P⬍.01) for the assigned grade 5 (176Nm), grade 4 (103Nm), grade 3 (67Nm), and grade 2 (19Nm). Healthy adults showed significant differences between grade 5 (212Nm) and grade 4 (120Nm) in mean torque (P⬍.05).
Conclusions: The supine MMT is a reliable and valid method with which to assess hip extension strength.
Key Words: Hip; Isometric contraction; Muscle, skeletal; Rehabilitation; Torque.
© 2004 by the American Congress of Rehabilitation
Medi-cine and the American Academy of Physical MediMedi-cine and Rehabilitation
A
LTHOUGH HIP EXTENSOR strength in the elderly has been identified as a primary predictor of walking ability,1,2 physical performance,3 and balance,4 assessment of the hip extensor strength is commonly overlooked. This oversight can be attributed to 2 reasons. The function of the hip extensors during walking is subtle, and the prone position required for conventional manual muscle testing of the hip extensors often is a major inconvenience.During walking, the hip extensors have a vital, stabilizing role at the onset of each stride. As body weight is dropped onto
the forward limb, the extensor muscles contract sharply to preserve upright stance by resisting forward fall of the pelvis and the trunk. The hip’s flexed position, before attaining the passive stability provided by full extension in midstance, cre-ates the demand.5 When this protective posture cannot be attained, the hip extensors must continue their support.5
In the standing position, weakness of the hip extensors causes the pelvis to fall forward. The trunk follows the pelvis whenever the patient cannot substitute with compensatory lor-dosis. The resulting forward trunk posture catches the clini-cian’s eye, whereas the role of the hip extensors is overlooked. Coexistence of inadequate spine flexibility and hip extensor weakness is a common finding in patients disabled by hip pathology, incomplete spinal cord injury, or spine fusions associated with paralysis.
Limitations in spine and hip mobility also make it difficult to assess hip extensor function. Lovett’s traditional manual mus-cle test (MMT) of the hip extensors requires that the patient be turned prone to include gravity as a grading factor.6Although easily accomplished in the era of childhood polio, when the procedure was developed, the presence of pain, contractures, and reduced mobility limits the applicability of this technique to today’s adult population with hip or spine impairment. Deprived of an antigravity testing position, clinicians have chosen muscle groups amenable to supine testing when relating muscle strength to function. Combined testing of the hip flex-ors and quadriceps is the approach used in the extensive Wom-en’s Health and Aging Study.7Correlations of hip flexor and quadriceps strength with a chair standing task were linear only with strengths below a threshold, and then accounted for less than 25% of the disability. When related as torque/body mass (Nm/kg), only knee extension served as a significant predictor of maximum walking speed.8 Improvements in quadriceps strength have also been reported in relation to reductions in gait deviations.
In contrast, studies that include the hip extensors demon-strate notably stronger correlations between muscle strength and lower-extremity function. The hip extensors were the “only significant independent predictor” of walking speed (r⫽.611,
R2⫽.37) and stride length (r⫽.590, R2⫽.35) when isokinetic
hip extension, knee extension, and ankle plantarflexion strength were correlated with walking speed in a group of elderly men at risk for falls.1(p1155) The hip extensors also registered the highest correlations with the Physical Performance Test3 (r⫽.69), Duncan Functional Reach Test4(r⫽.45), and walking speed (r⫽.76).2The functional significance of the hip exten-sors, as shown by the previously mentioned studies, indicates a need to modify the current means of manually testing hip extensor strength so that people intolerant of prone lying also could be adequately assessed.
Limited opportunities to assess patients who are unable to lie prone undoubtedly has encouraged other clinicians to estimate hip extension strength with the patient supine. And, there have been no criteria for grading the outcome. To rectify this limi-tation, documentation of the supine manual test was begun as a comparison with the traditional prone, antigravity muscle testing technique9,10 applied to 9 postpolio patients. Three
From the Pathokinesiology Laboratory, Rancho Los Amigos National Rehabilita-tion Center, Downey, CA.
Presented in part as the Shands Lecture, Orthopaedic Research Society, 1999, Los Angeles, CA.
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors(s) or upon any organization with which the author(s) is/are associated.
Reprint requests to Jacquelin Perry, MD, Pathokinesiology Laboratory, Rancho Los Amigos National Rehabilitation Center, 7601 E Imperial Hwy, Downey, CA 90242, e-mail: [email protected].
0003-9993/04/8508-8517$30.00/0 doi:10.1016/j.apmr.2003.09.019
patients could not lie prone. Among the other 6 subjects, the same strength grade was recorded by both the supine and prone techniques in 10 of the 12 tested limbs, representing an 83% agreement. Based on these encouraging results, we initiated a formal study.
The objective of the present study was to identify a reliable supine testing technique that differentiated at least 4 levels of hip extensor muscle strength that were grossly similar to the existing prone system. The biomechanical rationale for the supine test is that greater positions of hip flexion provide a mechanical advantage for the hip extensors, which significantly increases their force production.11Our first hypothesis was that maximum hip extensor muscle capability is displayed by the ability to lock the hip in full extension against a maximum flexor torque. The second hypothesis was that greater hip extensor weakness requires the advantage of increased hip flexion to lock the hip.
METHODS Participants
Subjects were recruited from Rancho Los Amigos National Rehabilitation Center and the surrounding Los Angeles area to participate in 1 of the 2 phases of the study: reliability testing or validity assessment. To determine the interrater reliability of the supine hip extensor test, 16 patients attending a regular postpolio clinic were recruited (table 1). Both lower limbs were assessed by each examiner. Pain prevented 1 subject from
completing the test with her right leg. As a result, reliability was calculated from the data on 31 limbs.
For objective validation and definition of the grading scale used in the test, 2 groups of subjects were recruited. Eighteen adults without known neurologic or limb pathology were re-cruited from the local Los Angeles area and were assigned to the no-pathology (NP) group (table 1). Only the left lower limb was tested in these subjects because it was more easily filmed from the standard location of our portable laboratory video camera. For the pathology (P) group, 26 persons with clinical indications of hip extensor weakness were recruited (table 1). The diagnoses of this group included postpolio syndrome (n⫽11), hip joint osteoarthritis (n⫽6), incomplete cauda equina injuries (n⫽4), and Guillain-Barre´ syndrome (n⫽2). Two subjects were impaired by generalized lower-extremity arthritic pain and 1 by previous ankle fractures. Although all patients were tested bilaterally, pain prevented 1 subject from completing the procedure with her right leg. As a result, P group data were from 51 limbs. This study was approved by the institutional review board at Rancho Los Amigos National Rehabilitation Center, and data were collected from November 1998 until April 2001. All subjects gave informed consent. Instrumentation
A chain suspended from the ceiling and fitted with a cable tensiometer was used to measure the force during isometric hip extension. The chain was fastened to the subject’s ankle with a padded leather cuff 10cm wide. The lever arm length of the test limb was hand measured with a calibrated metal tape to permit torque calculation (fig 1). All tests were videotaped, with the examiner taking care not to obstruct the lateral view of the test limb.
Procedure
Supine MMT. Clients were positioned supine on a firm surface and straight-leg raise range was assessed to determine if hamstring length was sufficient for them to assume the testing position and to obtain a sense of relaxed limb weight. The examiner then placed both hands under the heel (fig 2A), and the subject was asked to press the test limb into the mat while the examiner lifted the limb at least 90cm. The examiner also instructed the participant to keep the hip “locked” (ie, not allow the hip to flex). The subjects were not given any instruc-tions regarding the position of the contralateral limb.
Table 1: Subject Characteristics
Subject Group
Reliability
Study Validity Study Postpolio (n⫽16) No Pathology (n⫽18) Pathology (n⫽26) Age (y) 52⫾13.6 42.4⫾19.1 54.1⫾15.8 Mass (kg) NA 75.2⫾11.7 78.7⫾17.5 Men 9 11 11 Women 7 7 15 Limbs tested 31 18 51
NOTE. Values are mean⫾ standard deviation (SD) or as otherwise indicated.
Abbreviation: NA, not applicable.
Fig 1. Experimental setup for testing maximum isometric hip extension force using a ca-ble tensiometer (T) and lever arm measurement.
Grades were assigned according to predetermined criteria (table 2). The grading criteria were confirmed from the video records after the tests. Grade 5 (normal) was assigned if the subject was able to maintain the hip at neutral with the pelvis and back rising from the surface as a locked unit, while the leg was elevated by the examiner (fig 2B). Grade 4 (good) was indicated if the subjects allowed a small arc of hip flexion (not exceeding 30°) before they established a locked hip with the pelvis elevating from the table as the examiner lifted the leg (fig 2C). Grade 3 (fair) was given if the hip flexed throughout the test despite the examiner sensing a “good” amount of resistance. There was zero to minimal elevation of the pelvis (fig 2D). Full hip flexion with minimal resistance and no pelvic elevation was assigned a grade 2 (poor) (fig 2D). Full hip flexion with an absence of any resistance was assigned a grade 0 (absent) (fig 2D). Grade 1 (trace) was not used because the
supine posture, combined with body pressure through the but-tocks, obscured the palpation or visualization of a minimal muscle contraction.
To identify intertester reliability, 2 physical therapists, each unaware of the other’s finding, manually tested the subjects’ hip extensor strength bilaterally by using the supine technique. Testers performed their evaluations on the same day during the subject’s routine outpatient clinic visit. An adequate rest break was provided to minimize the effects of fatigue on perfor-mance.
Maximal isometric hip extension. For each limb, maxi-mum isometric hip extensor torque (newton meters) was cal-culated as the product of the maximum hip extensor muscle force (in newtons) multiplied by lever arm length (in meters). The extensor force was recorded with patients lying supine on a firm testing table. The pelvis was stabilized with a broad Velcro strap. A leather cuff was placed at the level of the ankle and the tested limb was elevated to 20° of hip flexion. A cable tensiometer, suspended from the ceiling by a chain, held the leg in place. The cable was aligned perpendicular (90° angle) to the limb (fig 1). Passive limb weight was obtained from the tensi-ometer reading while the relaxed limb was suspended by the cuff in the test position. Hip extension force was recorded as the subject pulled downward with maximum effort for 5 sec-onds. Tensiometer values were recorded digitally at a sampling rate of 2500Hz and averaged over .05-second intervals during the 5-second test. From these data, a moving window technique was used to determine the highest .25-second interval of force. Hip extension force equaled total extension force minus the passive limb weight previously determined with the cable tensiometer. The test lever arm was measured from the upper anterior margin of the trochanter (hip joint center) to the middle of the ankle cuff (fig 1).
Data Analysis
To determine the interrater reliability of the 2 examiners performing the supine hip extensor test, the statistic was calculated. The statistic provided a “chance-corrected”
mea-Fig 2. Supine hip extensor MMT. (A) Starting position for test. (B) Ending position for grade 5 (normal). Pelvis and back elevate as a locked unit while the leg is raised by the examiner. The hip maintains the fully extended, neutral po-sition throughout the test. (C) Ending position for grade 4 (good). Hip flexion occurs be-fore pelvis elevates while the examiner raises the leg. (D) Ending position for grades 3 (fair) and 2 (poor). Full eleva-tion of the limb to the end of the straight-leg raising range with no elevation of the pel-vis. Examiner feels “good” sistance for grade 3, little re-sistance for grade 2, and no active resistance for grade 0.
Table 2: Grading Criteria for the Supine Hip Extensor Test
Grade 5 (normal)
Hip locks in neutral (full extension) throughout the test.
Pelvis and back elevate as a locked unit as the leg is raised by the examiner.
Grade 4 (good)
A limited arc of hip flexion occurs before the pelvis and back elevate as a unit while the leg is raised by the examiner.
Grade 3 (fair)
Full flexion of the hip to the end of straight-leg raising range with little or no elevation of the pelvis.
Examiner feels “good” resistance throughout the test.
Grade 2 (poor)
Hip flexes fully with only minimal resistance felt by the examiner as the limb is raised. Examiner perceives that resistance exceeds that
due to leg weight. Grade 0
(absent)
Hip flexes fully with no active resistance felt by examiner as limb is raised.
Examiner perceives that resistance is due to leg weight only.
sure of agreement (ie, it mathematically excludes the incidence of chance agreements) for strength grades (5, 4, 3, 2, 0) assigned by the 2 therapists who independently tested the 31 limbs of persons with postpolio.
For validity assessment, peak torque producing capability was analyzed. The peak isometric hip extension torque was defined as the highest average value recorded during a .25-second sample of the 5-.25-second recording of maximum efforts. Each of the 69 tested limbs of persons in the P and NP groups was assigned to 1 of 4 strength grade groups (5, 4, 3, 2), according to the hip extensor’s response to the supine MMT. A 1-way analysis of variance was performed to compare torque data across the 4 strength grade groups. Significance testing was set at an ␣ level of .05. P values were corrected for multiple comparisons. All data were analyzed by using BMDP statistical software.a
RESULTS Reliability Testing
The repeated supine testing of hip extensor muscle strength resulted in a chance-adjusted measure of agreement of 82% between the grades independently assigned by the 2 physical therapists. This represents an excellent level of agreement in grades across the 31 limbs tested, according to guidelines presented by Landis and Koch.12Grading was identical for 27 of the 31 limbs, and the grades for the other 4 limbs were within a single level.
Validity Testing
The 18 limbs of people in the NP group showed 2 levels of hip extensor muscle strength (table 3). Grade 5 (normal strength) was accomplished by 16 limbs, whereas the other 2
Fig 3. Prone leg elevation with (A) pelvis free and (B) pel-vis restrained. Note the appar-ent decreased hyperextension range when lumbar lordosis and pelvic elevation are re-stricted.
were grade 4 (good). The mean hip extension torque was 212Nm for the group with grade 5 and 120Nm for the 2 subjects with grade 4 hip extensors (table 3). The difference in mean peak torque between these 2 strength levels was statis-tically significant (P⬍.05).
In the P group, hip extensor muscle strength of the 51 limbs was distributed among 4 grades (table 3). Five limbs had grade 5 extensor strength (normal) and 14 limbs were assigned grade 4 (good). The largest group, 22 limbs, displayed grade 3 (fair) strength. Ten were rated grade 2 (poor). The age of the subjects with grade 5 strength did not differ from that of subjects in the other groups.
Mean hip extension torques for the 4 groups were 176Nm (grade 5), 103Nm (grade 4), 67Nm (grade 3), and 19Nm (grade 2) (table 3). The differences in the mean torque levels for these 4 grades of hip extensor strength were significant at P less than .01. Although the mean torques for both grades 4 and 5 of the limbs evaluated in the P group were less than those of the NP group, the differences were not statistically significant.
DISCUSSION
Consistent with the principles of introducing a new proce-dure, the supine hip extensor test was assessed for reliability and validity. The 82% agreement between 2 therapists applying the supine test to the same 16 subjects identified a level of reliability that compared favorably with the other 2 reliability studies of hip muscle testing.13,14A 79% concurrence level was reported for 3 examiners testing the hip extensors of 4 sub-jects.13Repeat testing of hip flexor strength in 11 patients with a 2-day interval by 1 therapist yielded a 74% degree of agree-ment.14
Validity of the strength grades defined by the supine test was determined by isometric measurement of the subjects’ hip extensor muscles with a cable tensiometer. Among P group subjects, each grade registered statistically discrete differences in hip extensor strength (torque): grade 5 (175.6Nm), grade 4 (103.1Nm), grade 3 (66.7Nm), and grade 2 (19.1Nm).
The terminal hip postures (full extension, 30° flexion and no limit), which were used to differentiate grades 5, 4, and 3 of hip extensor muscle strength, showed the manner in which patients used the mechanical advantage of hip flexion to enhance their extensor muscle strength. The basis for this substitution rests on the findings of Waters et al,11who found that hip extensor muscle torque registered with the hip fully extended increased 41% with 15° of hip flexion and 112% with 45° of flexion. This represented an average increase of 1.9Nm per degree of flex-ion. For 20° of flexion, a 54% gain was calculated. During the supine MMT, patients with grade 5 hip extensor strength (176Nm) successfully resisted the examiner’s testing torque with their hip in full extension. In contrast, patients with grade 4 hip extensor strength (103Nm) required the mechanical ad-vantage of almost 30° of flexion to match the examiner’s challenge. The subjects with only grade 3 strength (67Nm)
displayed notable resistance but could never fully oppose the examiners’ lift force despite 60° of flexion. Hence, locking the hip in flexion rather than full extension is a sign of hip extensor weakness.
However, for the supine hip extension test to be accepted by clinicians there must be evidence that the grades are compara-ble to those obtained with the traditional prone MMT. We did not simultaneously test the same subjects in prone and supine positions because of the inconsistency in patient tolerance of the prone position. Instead, the supine data were compared with Beasley’s authoritative values for prone hip extensor grades.15 To accomplish this comparison, however, the supine test data had to be expressed in the same units that were used by Beasley.15 This required multiple adaptations. There were 4 significant differences in the measurement methods and data management to be addressed. One, Beasley recorded muscle strength as the force obtained with the cable tensiometer, whereas in the supine test study, force was multiplied by limb lever length and reported as torque (moment). In addition, Beasley used “force per body weight” (F/BW) as his basic strength unit to accommodate the wide range in the age of his subjects (age range, 5–29y).15The availability of simple force and body weight values for all subjects in the supine test study allowed recalculation of their strength data as F/BW. Two, Beasley also used a cable tensiometer to measure hip extensor strength with the knee fully extended, but the hip was in 45° of flexion. For the supine test, a 20° position was used (which more accurately mimicked weight acceptance limb posture). Based on Water’s data,11the hip force at 45° of flexion would be 139% greater than the force obtained at 20°. Applying this correction factor to Beasley’s data reduced the normalized force of his control group from .62 F/BW to .45 F/BW. Three, Beasley’s corrected force-to-body weight ratio (.45) still was significantly higher than the Rancho NP data (.37). This dif-ference showed that Beasley’s control (NP) group had greater strength per unit body weight than the Rancho control group. Consequently, another correction factor was applied to equal-ize the strengths of the 2 control groups. Four, the final calcu-lation followed Beasley’s presentation of the different levels of hip extensor strength as a percentage of the normal control group strength. The result was supine test values that were very similar to Beasley’s15standard (table 4).
The purpose in developing the supine hip extensor muscle test was not to replace the traditional prone technique but to provide a means of testing patients who otherwise would be inadequately evaluated because they cannot lie prone. The supine test, however, does offer some other advantages. Grades 5 and 4 can be differentiated by objective signs (hip position) rather than by relying on the examiner’s sense of resistance force. The ease of testing hip extensor strength is not dependent on spine or hip flexibility with the patient supine (fig 3). Significant hip flexion deformities, however, would make the grade 5 test position unavailable.
Table 3: Mean Peak (Nm) Isometric Hip Extension Torque per Muscle Grade for the NP and P Groups
Grade 2 (poor) Grade 3 (fair) Grade 4 (good) Grade 5 (normal)
NP group Torque* n⫽0 n⫽0 n⫽2 n⫽16
120.2⫾30.0 211.9⫾55.8
P group Torque† n⫽10 n⫽22 n⫽14 n⫽5
19.1⫾15.0 66.7⫾24.6 103.1⫾23.0 175.6⫾67.6
NOTE. Values are mean⫾ SD.
*Significant difference with grade 5⬎grade 4 (P⬍.05).
†
One final concern is the method of hip extensor testing. Full hip extensor strength involves 5 muscles—3 hamstring and the 2 prime hip muscles (gluteus maximus, adductor magnus). All 5 muscles are normally active during walking, but their partic-ipation varies with the subject’s pathology. Waters et al11 calculated the contribution of the hamstrings to be approxi-mately 35% of the total torque when the knee is fully extended and they found significant decreases with knee flexion. This situation requires continued use of a cable tensiometer to measure hip extensor strength. Substitutions with a static dynometer, for which the resistance lever arm extends only the length of the femur, would contribute to unpredictable changes in hamstring participation due to knee flexion.
CONCLUSIONS
First, MMT grading of the hip extensor muscles with the patient supine identifies strength levels similar to those identi-fied by the prone test. Second, 4 grades can be determined with the supine test: grade 5 (64%N), grade 4 (43%N), grade 3 (20%N), and grade 2 (10%N). Third, the supine muscle grad-ing postures reflect patients’ use of the mechanical advantage of hip flexion to supplement their hip extensor weakness. Last, the large, misleading arc of lumbar spine extension (fig 3) is avoided with the supine hip extensor test (fig 2B).
References
1. Burnfield JM, Josephson KR, Powers CM, Rubenstein LZ. The influence of lower extremity joint torque on gait characteristics in elderly men. Arch Phys Med Rehabil 2000;81:1153-7.
2. Salsich GB, Mueller MJ. Relationships between measures of function, strength and walking speed in patients with dia-betes and transmetatarsal amputation. Clin Rehabil 1997;11:60-7. 3. Reuben DB, Siu AL. An objective measure of physical function of elderly outpatients. The Physical Performance Test. J Am Geriatr Soc 1990;38:1105-12.
4. Duncan PW, Weiner DK, Chandler J, Studenski S. Functional reach: a new clinical measure of balance. J Gerontol 1990;45: M192-7.
5. Perry J. Gait analysis, normal and pathological function. Thoro-fare: Slack; 1992.
6. Lovett RW. The treatment of infantile paralysis. Philadelphia: Blakiston’s Son; 1916.
7. Ferrucci L, Guralnik JM, Buchner D, et al. Departures from linearity in relationship between measures of muscular strength and physical performance of the lower extremities: the Women’s Health and Aging Study. J Gerontol A Biol Sci Med Sci 1997;52:M275-85. 8. Rantanen T, Guralnik JM, Izmirlian G, et al. Association of
muscle strength with maximum walking speed in disabled older women. Am J Phys Med Rehabil 1998;77:299-305.
9. Hislop HJ, Montgomery J. Daniels and Worthingham’s muscle testing. 7th ed. Philadelphia: WB Saunders; 2002.
10. Kendall FP, McCreary EK, Provance PG. Muscles testing and function. 4th ed. Baltimore: Williams & Wilkins; 1993. 11. Waters RL, Perry J, McDaniels JM, House K. The relative
strength of the hamstrings during hip extension. J Bone Joint Surg Am 1974;56:1592-7.
12. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-74.
13. Agre JC, Magness JL, Hull SZ, et al. Strength testing with a portable dynamometer: reliability for upper and lower extremities. Arch Phys Med Rehabil 1987;68:454-8.
14. Wadsworth CT, Krishnan R, Sear M, Harrold J, Nielson DH. Intrarater reliability of manual muscle testing and hand-held dy-nametric muscle testing. Phys Ther 1987;67:1342-7.
15. Beasley WC. Quantitative muscle testing: principles and applica-tions to research and clinical services. Arch Phys Med Rehabil 1961;42:398-425.
Supplier
a. BMDP Statistical Software, Inc, 1440 Sepulveda Blvd, Los Ange-les, CA 90025.
Table 4: Comparison of Relative Strength of Hip Extensor Muscles (Defined as % Normal Control Group) Between
Supine Study Data and Beasley Data15 Grade 5 (normal) Grade 4 (good) Grade 3 (fair) Grade 2 (poor)
Rancho’s supine torque at
20° of flexion 83 49 32 9
Beasley’s F/BW at 45° of
flexion 65 43 20 5
Rancho supine corrected for definition, position,
