Hamstring Muscle Strain Treated by Mobilizing the Sacroiliac Joint

MICHAEL T. CIBULKA, STEVEN J. ROSE, ANTHONY DELITTO, and DAVID R. SINACORE

The purpose of this study was to compare the effectiveness of two types of treatment of hamstring muscle strains. Twenty patients with hamstring muscle strains were assigned randomly to an Experimental Group (n = 10) or a Control Group (n = 10). Peak torque production of the quadriceps femoris and hamstring muscles and hamstring muscle length were measured before and after treatment. The hamstring muscles of the Experimental and Control groups were treated with moist heat followed by passive stretching. The Experimental Group also received manipulation of the sacroiliac joint. The change in hamstring muscle peak torque was significantly greater for the Experimental Group than for the Control Group (p < .005). No significant differences existed between the two groups in either quadriceps femoris muscle peak torque or hamstring muscle length. The results of this study suggest a relationship between sacroiliac joint dysfunction and hamstring muscle strain.

Key Words: Manipulation, orthopedic; Physical therapy; Sacroiliac joint; Sprains and strains.

Many different factors have been implicated as possible causes of hamstring muscle strain.1-3 Despite such implica-tions, however, minimal clinical research has been conducted on its causes. Liemohn reported that a lack of hamstring muscle flexibility may be a predisposing factor in the devel-opment of hamstring muscle strain.1 Other reported factors include a strength imbalance between the left and right ham-string muscles,1 a strength imbalance between the quadriceps femoris and hamstring muscles,2 and sports that involve run-ning or jumping.3 Slocum and Bowerman have postulated that pelvic tilt is an important factor in postural control during running.4 Klein and Roberts believe that an anterior tilt of the pelvis can overstretch the hamstring muscles and can be a definite cause of hamstring muscle strain.5 To date, sacro-iliac joint dysfunction has not been implicated as a cause of hamstring muscle strain. We have noted a high correlation between hamstring muscle strains and an anterior tilt of the innominate bones resulting from sacroiliac joint dysfunctions. We have observed that those athletes who had hamstring muscle strain and were treated by mobilizing the sacroiliac joint were able to regain muscle function and return to activity sooner than those athletes who were treated with other, con-servative methods.

The purpose of this study was to compare the effects of manipulating the sacroiliac joint with those of a conventional method of treatment (ie, moist heat and stretching) on ham-string muscle strains. Because of our previously observed relationship between sacroiliac joint dysfunction and ham-string muscle strain, we expected that those subjects who were treated by manipulating the sacroiliac joint would regain muscle function (flexibility and peak torque) to a greater degree than those subjects who were treated solely with a conventional method consisting of moist heat and passive scretching.

METHOD

Subjects Twenty subjects (18 male subjects and 2 female subjects),

all with hamstring muscle strains, participated in the study. All of the subjects gave their informed consent before partic-ipating in the study. Their mean age was 24.5 years (range, 12-35 years). The average amount of time between onset of their injuries and their inclusion in the study was 9.5 days, with a range of 2 to 21 days. All of the subjects were patients at the St. Louis Sports Medicine Clinic. The diagnosis of hamstring muscle strain in this study was confirmed by the presence of 1) pain or ecchymosis, or both, localized to the involved hamstring muscle, 2) pain on resistive isometric knee flexion of the involved hamstring muscle, and 3) pain on passive stretching of the involved hamstring muscle. All testing was conducted at the St. Louis Sports Medicine Clinic.

The subjects were assigned randomly to an Experimental Group (n = 10) or a Control Group (n = 10). All of the subjects in the study had evidence of sacroiliac joint dysfunc-tion. Sacroiliac joint dysfunction was defined operationally as pelvic asymmetry between the left and right innominates,6 a

Mr. Cibulka is a physical therapist, St. Louis Sports Medicine Clinic, 14377 Woodlake Dr, Suite 311, Chesterfield, MO 63017 (USA).

Dr. Rose is Director and Associate Professor, Program in Physical Therapy, Washington University School of Medicine, and Co-Director, Department of Physical Therapy, Irene Walter Johnson Rehabilitation Institute, St. Louis, MO 63110.

Mr. Delitto and Mr. Sinacore are Instructors, Program in Physical Therapy, Washington University School of Medicine.

This article was submitted November 13, 1984; was with the authors for revision 30 weeks; and was accepted January 9, 1986.

1220 PHYSICAL THERAPY

RESEARCH

positive standing-flexion test,7 and a positive prone knee-flexion test (prone leg-length test).7 Although no studies have demonstrated the reliability of the standing-flexion test and the prone knee-flexion test, we have found these tests to be reliable in the clinic. In a preliminary study, two experienced clinicians assessed 25 patients using the prone knee-flexion test and the standing-flexion test and agreed on the test results of 22 of the 25 patients. The amount of tilt of the left and right innominates (the angle of inclination) was measured in degrees, and a four-degree or greater difference between the left and right innominate bones was defined as pelvic asym-metry.6 Pelvic calipers were used to measure the angle of inclination. This measure has been shown to be reliable when all testing is conducted on the same day (r = .93).8 (Briefly, this method involves placing the tip of one of the calipers on the anterior-superior iliac spine and the other on the ipsilateral posterior-superior iliac spine. An apparatus consisting of a protractor with a plumb line was attached to the calipers so that a measurement of the angle of inclination in degrees from the horizontal plane could be obtained.)

Procedure

We determined hamstring muscle length by measuring the amount of passive knee extension with the ipsilateral hip joint positioned at 90 degrees of hip flexion. The opposite lower extremity remained in its anatomical position. Full extension of the knee was defined as 180 degrees of extension from the horizontal plane.

Immediately after the assessment of the subject's hamstring muscle length, the peak torques of both the involved and uninvolved hamstring muscles and of both quadriceps femoris muscles were measured. A Cybex II isokinetic dynamome-ter* was used to measure peak torque. This dynamometer has been shown to be reliable with an interrater correlation coef-ficient of .95, especially when all testing is completed in one day.9 The subject's knee-joint axis was aligned with the lateral joint line of the knee. The subject's trunk and thighs were stabilized in place with straps to prevent excessive motion. We set the dynamometer at 60/sec, and the subjects were allowed to perform warm-up exercises for five minutes before testing. The subjects were instructed to start the warm-up exercises with submaximal contractions and to increase grad-ually to maximal contractions. The subjects then were asked to perform three maximal extensions followed immediately by flexion contractions within their limits of pain. All of the subjects stated that the isokinetic testing produced only a minimal amount of pain in their hamstring muscles. We believe that even if the isokinetic testing caused enough pain to limit peak torque production, the dynamometer still would be an appropriate measure of the amount of pain the subjects were experiencing. The highest value of the three maximal contractions was recorded as the peak torque measure. A damper setting of 2 was used to minimize the initial overshoot that can occur with the Cybex II dynamometer.10

The Experimental Group subjects received moist heat to the involved hamstring muscle for 20 minutes. This treatment was followed by three 2-minute repetitions of passive stretch-ing of the involved hamstring muscles. After the heat and

stretching treatments, one manipulative technique was ap-plied to the sacroiliac joint. This technique is described in detail by Erhard and Bowling.11 Briefly, this technique is performed while the subject is in a supine position and the lumbar spine is sidebent so that its concavity and the side of the injured hamstring muscle are away from the therapist. We instructed the subject to clasp his hands together behind his neck. The therapist (M.T.C.) then threaded one arm through the subject's clasped hands, rotating the subject to-ward him, and placed his free hand on that part of the subject's anterior-superior iliac spine that was farthest away from him. The therapist manipulated the sacroiliac joint by pushing down on the subject's anterior-superior iliac spine while ro-tating the subject's upper body toward him. Immediately after this treatment session, the therapist reassessed the subject's sacroiliac joint dysfunction, left and right quadriceps femoris and hamstring muscle peak torque, and left and right ham-string muscle length.

All of the subjects in the study had a sacroiliac joint dysfunction. It was determined by the prone knee-flexion test and by determining which innominate was posterior when measuring the angle of inclination with the pelvic calipers. After only one treatment involving the manipulative tech-nique, all subjects in the Experimental Group exhibited a symmetrical pelvis and negative results on the standing-flex-ion and prone knee-flexion tests.

The Control Group was treated identically to the Experi-mental Group except that no manipulative technique was applied to the sacroiliac joint. After the treatment session was completed, the therapist reassessed the subject's quadriceps femoris and hamstring muscle peak torques and hamstring muscle length. All of the subjects in the Control Group, after treatment, still exhibited an asymmetrical pelvis and positive results on the standing-flexion and prone knee-flexion tests.

Data Analysis

The data were analyzed using an analysis of covariance (ANCOVA),12 We used the ANCOVA to determine whether significant differences existed between the Experimental and Control groups in hamstring muscle length, quadriceps fem-oris muscle peak torque, and hamstring muscle peak torque. We also used the ANCOVA to adjust the posttest values to correct for initial differences in the pretest values.

RESULTS

We found that the Experimental and Control groups were significantly different when we compared their involved ham-string muscle peak torques (F = 12.66; df= 1,17; p < .005). Table 1 contains a summary of the ANCOVA results for hamstring muscle peak torque. Table 2 summarizes the means, standard deviations, and changes of hamstring muscle peak torque in foot-pounds at 60o/sec. The adjusted mean for hamstring muscle peak torque of the Experimental Group was 49 ft.lb and the adjusted mean for hamstring muscle peak torque of the Control Group was 42 ft.lb. We found no significant differences between the Experimental and Control groups in either quadriceps femoris muscle peak torque (Tab. 2) or hamstring muscle length (Tab. 3). * Cybex, Div of Lumex, Inc, 2100 Smithtown Ave, Ronkonkoma, NY 11779.

Volume 66 / Number 8, August 1986 1221

TABLE 1 Analysis of Covariance Results for Adjusted Posttest Hamstring Muscle Peak Torque Using Pretest Hamstring Muscle Peak Torque as the Covariate

Source

Group Error TOTAL

df

1 17 18

SS

286.69 384.95

MS

286.69 22.64

F

12.66 P

.005

TABLE 2 Means and Standard Deviations of Quadriceps Femoris and Hamstring Muscle Peak Torque at 60/seca

Group

Control Pretest Posttest

Experimental Pretest Posttest

Hamstring Muscle Peak

Torque

46.0 46.4

37.6 45.7

s

19.08 17.47

20.53 22.70

Quadriceps Femoris Muscle

Peak Torque

109.0 114.8

94.9 92.7

s

24.80 30.90

48.57 50.88

TABLE 3 Means and Standard Deviations of Hamstring Muscle Lengtha (180 Equals Full Knee Extension with the Hip Flexed to 90)

Group

Control Pretest Posttest

Experimental Pretest Posttest

132.6 144.6

140.0 155.0

s

17.1 16.7

16.8 14.2

DISCUSSION

The results of our study raise the following questions: 1) What mechanism accounted for the increase in hamstring muscle peak torque of the subjects in the Experimental Group? 2) Does failure of the musculotendinous unit occur first in the connective tissue or in the muscle? 3) Can the biomechanical functioning of other musculotendinous units attached to the pelvis also be disrupted by a sacroiliac joint dysfunction (eg, rectus femoris muscle strain or patellar ten-donitis)? 4) What adaptations take place in the hamstring muscles? 5) What adaptations occur in the connective tissue after the passage of time?

The mean hamstring muscle peak torque of the subjects in the Experimental Group increased after only one treatment. Their gain in hamstring muscle peak torque was significantly greater than that of the subjects in the Control Group (p < .005). The gain in hamstring muscle peak torque after one treatment may have resulted because the treatment reduced the stress on the subject's injured hamstring muscle.

A sacroiliac joint dysfunction creates an anterior tilt of the innominate on one side and a posterior tilt of the innominate on the opposite side. Manipulating the sacroiliac joint reduces both the anterior and the posterior tilt of the innominates. We believe that the stress on the hamstring muscle may have been decreased by restoring the normal relationship between the innominates. In all of our subjects we always have seen an anterior tilt of the innominate on the side of the hamstring muscle strain. An anterior tilt moves the origin of the ham-string muscle, the ischial tuberosity, farther from its insertion (Figure). An anterior tilt elongates the entire hamstring mus-culotendinous unit. Correcting the sacroiliac joint dysfunction may have reduced the length of the hamstring muscles. Per-haps, the subject could produce a greater hamstring muscle peak torque after manipulation than before manipulation because the hamstring muscle's normal resting length was restored.

That torque gains were recorded after the hamstring muscle was shortened does not support the concept of the length-tension curve. According to the length-tension curve concept, shortening the hamstring musculotendinous unit reduces a muscle's ability to develop tension. After reducing the ham-string muscle length, however, we recorded a greater peak torque for the hamstring muscles. The gain in hamstring muscle peak torque, therefore, could not have resulted be-cause of a change in the length-tension curve.

Figure. The relative difference in hamstring muscle length of a normal innominate (solid line) and an anteriorly tilted innominate (interrupted line).

a Peak torque in foot-pounds.

a Measured in degrees.

1222 PHYSICAL THERAPY

RESEARCH

Although we found a significant increase (p < .005) in the peak torque of the involved hamstring muscles of the Exper-imental Group subjects, no significant difference was found in hamstring muscle flexibility between the two groups. Be-cause the subjects' hamstring muscles were stretched in both the Control and the Experimental groups, we anticipated that we might not find a significant difference between the two groups. Further study is necessary to determine the effective-ness of this treatment on hamstring muscle length.

A sacroiliac joint dysfunction may be an important precip-itating factor in the development of hamstring muscle strain. This factor, however, does not preclude the involvement of other factors that also may cause hamstring muscle strains, such as inflexible hamstring muscles,1 weakness of the ham-string muscles,1 muscle strength imbalances between the quadriceps femoris and the hamstring muscles,2 abnormal stress from sports that involve running or jumping,3 and tight hip flexor muscles that cause a bilateral anterior pelvic tilt.5 Hamstring muscle strain probably is precipitated by one or more of these factors.

One limitation of this study is the use of change scores in its design. A change score is defined as the difference between the scores of the pretest and the posttest. The use of change scores does not affect the internal validity of an experiment, but may reduce the external validity. A pretest can make the subject more sensitive to the experimental manipulation.13 Consequently, studies that use pretests may only generalize to those groups that have been pretested. Inferences based on studies that use pretests, therefore, cannot be extrapolated to sample groups that are not pretested. The best approach to minimize the undesirable effects of a pretest is to use an ANCOVA.14 An ANCOVA adjusts the posttest values to correct for initial differences in pretest values and, thus, eliminates the pretest influence on the posttest scores. If a pretest is used, the subjects always should be assigned ran-domly, and the data should be analyzed with an ANCOVA or a multiple regression analysis.14

Because the testing was conducted by one person and was not conducted blindly, the results of our study may reflect the motivation of the researcher rather than the efficacy of the treatment. Because this study was conducted in a clinical setting as part of routine treatment procedures, it precluded the use of a double blind design.

The results of our study suggest that a relationship exists between sacroiliac joint dysfunction and hamstring muscle strain. The side on which the hamstring muscle strain devel-oped was always the side of the anterior tilt of the innominate; hamstring muscle strain never developed on the side of the posteriorly tilted innominate. Further studies are needed to determine whether we can predict the occurrence of hamstring muscle strains by assessing sacroiliac joint dysfunction. Ther-apists should recognize the importance of evaluating the sacroiliac joint and its effect on the two-joint musculature of the thigh.

CONCLUSIONS

The results of our study support our hypothesis that patients with hamstring muscle strains who are treated by correcting a sacroiliac joint dysfunction have a greater increase in peak torque in their injured hamstring muscles than those patients whose sacroiliac joints are not manipulated. Sacroiliac joint

dysfunction may predispose an athlete to muscle strain. Phys-ical therapists should examine the pelvis for possible dysfunc-tion in athletes who have muscle injuries of the thigh.

REFERENCES 1. Liemohn W: Factors related to hamstring strains. J Sports Med Phys

Fitness 18:71-75, 1978 2. Burkett LN: Causative factors in hamstring strains. Med Sci Sports 2:39-

42, 1970 3. Regua RK, Garrick JG: Injuries in interscholastic track and field. The

Physician and Sportsmedicine 9(3):42-49, 1981 4. Slocum DB, Bowerman W: The biomechanics of running. Clin Orthop

23:39-45, 1962 5. Klein KK, Roberts CA: Mechanical problems of marathoners and joggers

and solution. In Landry F, Orban WAR (eds): Sports Medicine, Medicine du sport. Miami, FL, Symposia Specialists, 1976, p 210

6. Pitkin HC, Pheasant HC: Sacroarthrogenetic telagia: A study of sacral mobility. J Bone Joint Surg 18:365-374, 1936

7. Mitchell FL, Moran PS, Pruzzo NA: An Evaluation and Treatment Manual of Osteopathic Muscle Energy Procedures, ed 1. Valley Park, MO, Mitchell, Moran, and Pruzzo, Associates, 1979

8. McClure JM, Mayhew T, Norton BJ: Validity and Reliability of Two Postural Measures: Pelvic Tilt and Femoral-Tibial Angle. Read at the Fifty-Ninth Annual Conference of the American Physical Therapy Association, Kansas City, MO, June 14-18, 1983

9. Johnson J, Siegel D: Reliability of an isokinetic movement of the knee extensors. Research Quarterly 49:88-90, 1978

10. Sinacore DR, Rothstein JM, Delitto A, et-al: Effect of damp on isokinetic measurements. Phys Ther 63:1248-1250, 1983

11. Erhard R, Bowling R: The recognition and management of the pelvic component of low back and sciatic pain. Bulletin of the Orthopaedic Section, American Physical Therapy Association 2(3):4-15, 1977

12. Ferguson GA: Statistical Analysis in Psychology and Education, rev ed 5. New York, NY, McGraw-Hill Inc, 1976

13. Campbell DT, Stanley JC: Experimental and Quasi-Experimental Designs for Research. Skokie, IL, Rand McNally & Co, 1966

14. Kerlinger FN: Foundations of Behavioral Research, ed 2. New York, NY, Holt, Rinehart & Winston General Book, 1973

Volume 66 / Number 8, August 1986 1223