Knee Joint Stiffness in Individuals With and Without Knee Osteoarthritis: A Preliminary Study

Carol A. Oatis, PT, PhD 1 Edward F. Wolff, PhD 2 Sandra K. Lennon, DPT, BS 3

Study Design: Descriptive, case-matched comparison. Objectives: To compare the knee joint stiffness and damping coefficients of individuals with knee osteoarthritis (KOA) to those of age- and gender-matched individuals without KOA. A secondary purpose was to investigate relationships between these coefficients and complaints of stiffness in individuals with KOA. Background: KOA is a leading cause of disability, and stiffness is a common complaint in individuals with KOA. Yet the most common method of assessing knee joint stiffness is through a self-report questionnaire. Methods and Measures: Stiffness and damping coefficients at the knee were calculated in 10 volunteers (mean age SD, 64.1 15.5 years) with KOA and compared to coefficients from ageand gender-matched individuals without KOA, collected in a previous study (mean age SD, 62.1 13.9 years). Stiffness and damping coefficients were calculated from the angular motion of the knee during a relaxed oscillation. Spearman correlation coefficients were calculated between stiffness and damping coefficients and WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index) scores for subjects with KOA. Results: Independent 2-tailed t tests revealed significantly larger damping coefficients (P = .035) among those with KOA (95% CI, 0.10-2.32 Nm s/rad). Spearman rank correlations revealed a significant positive relationship (r = .85, P = .003) between the damping coefficient and the stiffness subscore of the WOMAC. Conclusion: This study offers preliminary data demonstrating the feasibility of measuring stiffness and damping coefficients in individuals with KOA. Additionally, the damping coefficient is increased in people with KOA when compared to age- and gender-matched individuals without KOA. The damping coefficient appears to be associated with the complaints of stiffness reported by the WOMAC. J Orthop Sports Phys Ther 2006;36(12):935-941. doi:10.2519/jospt.2006.2320

Key Words: arthritis, pendulum test, tibiofemoral joint, WOMAC

steoarthritis is the most common form of arthritis and approximately 6% of adults 30 years old or older report symptomatic knee osteoarthritis (KOA).4 Prevalence increases with age.9 KOA is also a leading cause of chronic disability.3,6 Because the prevalence of KOA is expected to increase substantially with the aging of the baby boomers, it is
Professor, Department of Physical Therapy, Arcadia University, Glenside, PA. Associate Professor, Department of Computer Science and Mathematics, Arcadia University, Glenside, PA. 3 Staff Physical Therapist, Easter Seals, Levittown, PA. This study was completed without financial support and was approved by the Committee for the Protection of Research Subjects (COPRS) at Arcadia University (formerly known as Beaver College). Address correspondence to Carol A. Oatis, Department of Physical Therapy, Arcadia University, 450 S. Easton Road, Glenside, PA 19038. E-mail: oatis@arcadia.edu
2 1

important to understand the factors associated with KOA that may contribute to disability. Stiffness is a common complaint in individuals with osteoarthritis. It is 1 of 6 criteria used in the clinical diagnosis of KOA1 and may contribute to the disability associated with KOA. Yet despite the prevalence of KOA in the population and the consistency of the complaint of stiffness, there are few clinically feasible ways to assess a patients complaint of stiffness. The most common method of assessing this variable is through the self-report tool, Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC).2 The WOMAC consists of 24 separate questions, 5 addressing pain, 17 addressing function, and 2 addressing stiffness. Although this tool has been validated and is used consistently to assess treatment outcomes, it is limited in its ability to directly address an individuals specific complaints of stiffness. The pendulum test described by Wartenberg15 has been used to assess spasticity in neurologically impaired individuals by counting the number of swings that a relaxed limb undergoes when dropped from a suspended position. We have used a biomechanical model of the knee to characterize its stiffness and damping
935

RESEARCH REPORT

Journal of Orthopaedic & Sports Physical Therapy

coefficients using the pendulum test procedure.11 Stiffness is operationally defined as the ratio of the change in torque to a change in knee angle expressed in Newton-meters/radian (Nm/rad). The damping coefficient reflects the time-dependent nature of the deformation expressed in Newton-metersecond/radian (Nm s/rad). A similar approach has been used at the elbow in healthy subjects.10 Others have used the test to compare stiffness in individuals with and without patellofemoral joint syndrome.7 The purpose of this study was to use the biomechanical model to quantify the knee joint stiffness and damping coefficients in individuals with KOA and to compare these values to data previously collected from age- and gender-matched subjects without KOA. We hypothesized that the stiffness and damping coefficients in individuals with KOA would be larger than age- and gender-matched individuals without KOA. Additionally, we examined the correlations between stiffness and damping coefficients and WOMAC scores in those subjects with KOA.

TABLE 1. Repeated measures of stiffness and damping coefficients in 6 knees Day 1 Stiffness (Nm/rad) 8.25 5.81 21.07 19.92 13.64 11.29 Day 2 Stiffness (Nm/rad) 7.10 6.26 21.19 19.70 12.41 10.96 Day 1 Damping (Nm s/rad) 0.88 0.83 2.98 2.63 0.13 0.24 Day 2 Damping (Nm s/rad) 0.67 0.85 2.39 2.88 0.17 0.19

METHODS
The model used to quantify the stiffness and damping coefficients is described in detail elsewhere.11 It models the knee joint as a rotational spring with a damper joining a rigid thigh and leg-foot segment. Anthropometric measurements of the foot and leg were used to determine the moment of inertia of the leg-foot segment. The premise of the model is that stiffness and damping coefficients can be determined by tracking the displacement of the leg-foot segment during relaxed oscillations. The Appendix contains the equations used to calculate the stiffness and damping coefficients.

data from only 1 knee of the other subject. We used a coin flip to choose which knee to include. The data from the subjects were paired with data from ageand gender-matched subjects without osteoarthritis, collected in a previous study (Table 1).11 After finding subjects in the original study with matching gender, the subjects with the closest age were identified. If there was more than 1 subject with the same age match, the first subject who participated was chosen.

Instrumentation
The motion of the knee was monitored using a 3-camera computerized video system with ExpertVision Version 3.3 software (Motion Analysis Corporation, Santa Rosa, CA). Motion data were collected at 60 Hz.

Procedure
Prior to data collection, each volunteer was instructed in the purpose of the study and the test procedures. Each volunteer then read and completed an informed consent. Next, the examiner performed a physical examination of each knee tested, and each subject with KOA completed a WOMAC self-report questionnaire. Anthropometric measures of the leg and foot were recorded and used to calculate the moment of inertia with respect to the knee to be used in the calculations of stiffness and damping coefficients. With the subject seated on an elevated table and the untested leg supported on a stool, retroreflective markers were applied to the skin overlying the acromion, greater trochanter, lateral epicondyle, and lateral malleolus of the test limb. The subject was instructed to sit upright and was positioned so that the test limb hung unsupported with the knee in 75 of flexion. A pillow was used as needed to support the subjects back. For each trial, the examiner pulled the relaxed knee into maximum extension without lifting the thigh from the table or stretching the hamstrings. Maximum extension was determined by end feel. The

Subjects
Eleven volunteers with KOA were recruited from the university community and surrounding neighborhood. Inclusion criteria were age greater than 20 years, radiological evidence of KOA in the painful knee, and moderate knee pain within the last 4 weeks (3 to 6 on a 10-point visual analog scale). Subjects were excluded if they had uncontrolled cardiac disease, a joint arthroplasty of the painful knee, paresis of either lower extremity, or severe osteopenia, or if they were currently receiving physical therapy. Eight of the subjects with KOA, including 1 subject with bilateral KOA, participated in a reliability assessment. Data from all 9 knees were assessed twice within a 2-week period and the 2 assessments were compared. Eleven subjects participated in the comparison and correlation study. Two subjects had bilateral KOA. Good data were available from only 1 knee for 1 subject. To compare coefficients in individuals with KOA to those without KOA, we included
936

J Orthop Sports Phys Ther Volume 36 Number 12 December 2006

tester passively extended the knee to the point at which any passive resistance to extension was felt. The tester then flexed the knee approximately 5 to eliminate any tension from the hamstrings. This position was the start position for the test. The examiner instructed the subject to relax the limb. Once relaxed, the examiner released the limb and allowed the knee to fall into flexion, oscillating in flexion and extension until it came to a stop. Knee motion was collected continuously for 5 seconds and recorded by the Motion Analysis system. Because the original study,11 as well as other studies,7,8 revealed that muscle activity could be identified by an interruption of the smoothly decaying sinusoidal motion of the leg, we did not use EMG data to confirm muscle relaxation in the current study. This study was approved by the Committee for the Protection of Research Subjects at Arcadia University (formerly known as Beaver College).

Data Analysis
The angular motion of the knee during the relaxed oscillations was used to calculate stiffness and damping coefficients using calculations reported in the Appendix and described elsewhere.11 Intraclass correlation coefficients (ICC2,1) were used to assess reliability of the stiffness and damping coefficients in people with KOA on 2 different test days within a 2-week period. An independent 2-tailed t test for samples with unequal variance (Welch-Satterthwaite t test) was used to compare the stiffness and damping coefficients from the subjects with KOA to coefficients from age- and gender-matched subjects without KOA contained in the database of 96 subjects who participated in a previous study.11 We also compared the data using the distribution-free Mann-Whitney U test. We used Spearman rank correlations to assess the relationships between the stiffness and damping coefficients and WOMAC scores.

was 64.1 (14.7) years and 62.1 (13.2) years for the matched controls. Subjects were not matched on size or body type. However, the moments of inertia were based on leg length and girth. These are reported in Table 2. There was no statistical difference in moment of inertia between the groups (independent t test, P .30). The stiffness and damping coefficients from the 10 individuals with KOA and the matched subjects without KOA are reported in Table 3. Independent 2-tailed t tests for groups with unequal variance revealed significantly larger damping coefficients (P = .035) among those with KOA (mean difference, 1.21; 95% CI of the difference, 0.10 to 2.32 Nm s/rad). There was no statistically significant difference in stiffness (P = .18; mean difference, 3.88; 95% CI of the difference, 2.11 to 9.87 Nm/rad). However, the confidence interval for stiffness and the sample size suggest the possibility of a type II error. Analysis of the stiffness data revealed a moderate effect size (d = 0.63). Power analysis revealed that a sample size of 50 individuals with KOA would be needed to reveal a difference at the .05 level with an 80% power. The Mann-Whitney U test results were consistent with those from the t test with P values of P = .02 and .21 for damping and stiffness, respectively. Spearman rank correlations revealed a statistically significant (P = .003) positive correlation (r = .85; 95% CI, 0.44 to 0.97) between the damping coefficient and the stiffness subscore of the WOMAC (Figure). One subject failed to answer all questions on the WOMAC and was not included in the analysis. No other correlations between coefficients and
180 160

RESEARCH REPORT

WOMAC Stiffness

140 120 100 80 60 40 20

RESULTS
Although 9 knees were originally entered into the reliability study, only 6 knees had complete data on both days. If a knee did not complete 3 full decaying oscillations, the stiffness and damping coefficients could not be calculated. Although all of the individuals tested appeared to complete the tests successfully, later analysis revealed that 3 individuals had complete data on only 1 of the 2 days. Assessment of reliability from the remaining knees revealed excellent reliability with ICCs2,1 greater than 0.97 for both stiffness and damping coefficients (Table 1). Tests from 1 subject with unilateral KOA did not yield acceptable oscillations for analysis, so 10 subjects with KOA were included in the comparison with controls. The mean age (SD) of the subjects with OA

Damping (Nm s/rad)

FIGURE. Relationship of the damping coefficient and the stiffness subscore of the WOMAC. Sample includes 9 subjects because 1 individual did not respond to all questions of the WOMAC. Spearman rank correlation equals 0.85 (P = .003).
937

J Orthop Sports Phys Ther Volume 36 Number 12 December 2006

TABLE 2. Age, gender, and moments of inertia of subjects with knee osteoarthritis (KOA) and matched controls. Gender-Matched Subjects Age (y) 63 61 42 39 69 74 79 57 58 79 62.1 13.9 Moment of Inertia for Subjects With KOA (kg-m2) 0.554 0.535 0.263 0.582 0.422 0.263 0.293 0.477 0.356 0.589 0.43 0.13 Moment of Inertia for Matched Subjects (kg-m2) 0.603 0.309 0.450 0.505 0.489 0.242 0.261 0.261 0.267 0.379 0.38 0.13

KOA Subjects Gender M M F M M F F M F M Mean SD

KOA Subjects Age (y) 64 63 44 38 70 81 84 57 59 81 64.1 15.5

TABLE 3. Stiffness and damping coefficients in subjects with knee osteoarthritis (KOA) and in matched controls. Damping coefficients differed significantly between those with and without KOA (P = .035). KOA Stiffness Coefficient (Nm/rad) 8.25 5.81 3.58 19.92 13.64 11.29 17.73 17.96 16.36 31.45 14.60 8.08* *Mean SD. Control Stiffness Coefficient (Nm/rad) 15.97 9.34 12.11 13.29 13.08 7.82 5.41 9.52 8.51 12.13 10.72 3.14* KOA Damping Coefficient (Nm s/rad) 0.88 0.83 0.32 2.63 0.13 0.24 4.89 1.46 3.02 0.63 1.50 1.55* Control Damping Coefficient (Nm s/rad) 0.30 0.48 0.26 0.31 0.30 0.25 0.24 0.36 0.11 0.30 0.29 .09*

WOMAC scores were statistically significant. There was no correlation between the stiffness and damping coefficients (Spearman correlation, r = 0.26; P = .5).

DISCUSSION
This study applied a novel assessment tool to individuals with KOA to quantify the stiffness and damping coefficients of the affected knee. In addition, we compared the coefficients from people with KOA to those without KOA. The reliability assessment revealed that the tests could be applied reliably to individuals with symptomatic KOA. Yet we were unable to collect adequate data from 3 individuals who participated in the reliability study. While the existing data exhibit excellent reliability, the inability to collect good data from all subjects limits the generalizability of our results. The use of EMG data and improved training of the research assistants to recognize bad data and to repeat the trial may improve the success of data collection. As we hypothesized, these preliminary data suggest that the damping coefficient in people with KOA is greater than those in people without KOA. We also
938

hypothesized that the stiffness coefficient would be greater in individuals with KOA. Our data did not demonstrate a statistically significant difference. However, examination of the stiffness data and the confidence interval suggests that a real difference may exist, but the sample size was too small to detect it. We can draw no definitive conclusions regarding the stiffness coefficient at this time. An important finding in our study is the apparent relationship between the damping coefficient and the complaints of stiffness reported by individuals with KOA. Heretofore there has been no means of assessing that complaint other than by the 2 relevant questions of the WOMAC. Our data suggest that the quantified pendulum test may provide direct measurement of one of the most common complaints in KOA, stiffness. In fact, the stiffness reported by so many individuals with KOA actually may reflect the viscoelastic property, damping, measured by the pendulum test. Because this approach has not been applied to people with knee OA before, there are no direct comparisons available from the literature. Wright and

J Orthop Sports Phys Ther Volume 36 Number 12 December 2006

Johns16 found an increase in stiffness and damping coefficients in individuals with rheumatoid arthritis. However, these authors used a torque displacement method to assess the coefficients, which may not be comparable to the relaxed-oscillation method. Wartenberg15 and Olgiatti et al12 used the pendulum test to identify increased stiffness in individuals with neurological disorders. However, these authors used a more qualitative assessment of stiffness based on the number of swings that occurred and the shape of the oscillation. Stein et al14 used a method similar to our approach and found no difference in stiffness and damping coefficients in 9 neurologically impaired individuals, as compared to 9 neurologically intact subjects. However, the authors did not match gender for all subjects in the 2 groups, which is likely to have affected the comparisons. In healthy subjects, men consistently demonstrate larger stiffness and damping coefficients than women.10,11,16 Hamstra-Wright et al7 used a similar method to compare the stiffness coefficients between subjects with patellofemoral syndrome and age- and gender-matched controls. These authors found no difference in stiffness coefficients between the 2 groups, but did not report any comparison of the damping coefficients. In this study we have made no attempt to identify the source(s) of the increased damping coefficients seen in individuals with KOA. Our own working hypothesis is that this increase reflects the structural changes that occur in KOA, including changes in articular cartilage, thickening and stiffening of the supporting ligaments, and perhaps the associated inflammation.5,9 Our preliminary findings of a relationship between damping and complaints of stiffness support the concept that the damping coefficient, at least, reflects the changes in the viscoelastic components of the joint. It is also possible that the surrounding musculature contributes in someway. However, quadriceps weakness is a common finding in subjects with KOA.13 So it is less likely that quadriceps strength could explain the larger value in these coefficients in individuals with KOA. Our own unpublished data also suggest that small changes in the starting knee position (less than 15) and, hence, the stretch on the hamstrings do not affect the damping or stiffness coefficients. We have not investigated the effect body size and body mass index may have on the measures of stiffness and damping. Although our preliminary data show no difference in moments of inertia, which are functions of limb length and girth between the 2 groups, it is premature to conclude that stature and body mass index have no effect on stiffness or damping. Further research is needed to clarify the source of the increases in damping in individuals with KOA. This study is limited by its sample size. Further research is needed to confirm the current findings in a larger sample. In addition, 3 of the 8 individuals

initially enrolled in the reliability portion of the study did not have complete data and were not included in the analysis. Although the remaining data analyzed demonstrated excellent reliability and a previous study with young adult subjects free of knee dysfunction also demonstrated moderate to strong reliability,11 further investigation is needed to verify the applicability of the test among people with KOA. Incomplete data can result from either an interrupted decaying oscillation or from an overly damped swing that prevented 3 full oscillations. We did not analyze the bad data to determine the cause of the problem. However, either case may have arisen from undetected muscle activity. This problem might have been reduced by collecting EMG data during each data collection. Others have noted, as we did in our original study, that active muscle contraction alters the decaying swing of the limb and EMG data are unnecessary to detect muscle activity.7,8 Some subjects may also have so much damping that analysis is possible only by using 1 or 2 full oscillations. Additional analysis is needed to determine any effect of calculating stiffness and damping coefficients using 1 or 2 oscillations compared to the analysis based on 3 oscillations. Another limitation of our study is the fact that we relied on the subjects self-report of radiologic evidence of KOA. The subjects confirmed that they had seen a doctor, had a radiograph taken and were told that they had arthritis. Direct confirmation from the physician would have strengthened our findings. Additionally, information about the severity of the joint changes, such as joint space narrowing or a Kellgren-Lawrence radiologic score, would have allowed us to consider whether disease severity correlated positively with the stiffness and damping coefficients. We plan to address this question in a future study. An obvious question unanswered by the current study is the clinical importance of the stiffness and damping coefficients reported here. Hamstra-Wright et al7 suggest that a meaningful difference in stiffness is 1.42 Nm/rad. The difference between means of the individuals with KOA and the matched control subjects was 3.88 Nm/rad, although the standard deviations of the 2 groups were 8.08 and 3.14 Nm/rad, respectively. The greater variability seen in the subjects with KOA may reflect the variability of the disorder, and additional research is needed to determine if the stiffness and damping measures reflect the level of disease or dysfunction. Future studies are also needed to confirm the association we identified between the damping coefficient and the WOMAC stiffness subscore. Additionally, there is a need to determine what, if any, relationship exists between these measures and function in individuals with KOA. Finally, if stiffness and damping coefficients are to be
939

RESEARCH REPORT

J Orthop Sports Phys Ther Volume 36 Number 12 December 2006

useful clinical measures, they must demonstrate good responsiveness to changes in patients complaints.

CONCLUSION
This study offers preliminary data that demonstrate that it is feasible to measure the stiffness and damping coefficients in individuals with KOA. The damping coefficients are increased in people with KOA when compared to age- and gender-matched individuals without KOA and appear associated with complaints of stiffness as measured by the WOMAC. Further research with a larger sample is needed to verify these results and to identify the clinical significance of the findings.

ACKNOWLEDGEMENTS
The authors gratefully acknowledge the technical assistance of Melissa Adonizio, Jane Gillum, Laura Huber, and Cheryl Kousen, who assisted in data collection, and Jamie Iwanczweski for assistance in manuscript preparation. We are also indebted to Drs Emanuele Curotto and H. Stephen Huber for their valuable contributions to the conceptual model.

REFERENCES
1. Altman RD. Criteria for classification of clinical osteoarthritis. J Rheumatol Suppl. 1991;27:10-12. 2. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol. 1988;15:1833-1840. 3. Dawson J, Linsell L, Zondervan K, et al. Epidemiology of hip and knee pain and its impact on overall health status in older adults. Rheumatology (Oxford) . 2004;43:497-504.

4. Felson DT, Zhang Y. An update on the epidemiology of knee and hip osteoarthritis with a view to prevention. Arthritis Rheum. 1998;41:1343-1355. 5. Fishkin Z, Miller D, Ritter C, Ziv I. Changes in human knee ligament stiffness secondary to osteoarthritis. J Orthop Res. 2002;20:204-207. 6. Guccione AA, Felson DT, Anderson JJ, et al. The effects of specific medical conditions on the functional limitations of elders in the Framingham Study. Am J Public Health. 1994;84:351-358. 7. Hamstra-Wright KL, Swanik CB, Ennis TY, Swanik KA. Joint stiffness and pain in individuals with patellofemoral syndrome. J Orthop Sports Phys Ther. 2005;35:495-501. 8. Jennings AG, Seedhom BB. The measurement of muscle stiffness in anterior cruciate injuries -- an experiment revisited. Clin Biomech (Bristol, Avon). 1998;13:138140. 9. Jordan JM, Kington RS, Lane NE, et al. Systemic risk factors for osteoarthritis. In: Felson DT, conference chair. Osteoarthritis: new insights. Part 1: The disease and its risk factors. Ann Intern Med. 2000;133:635-646. 10. Lin CC, Ju MS, Huang HW. Gender and age effects on elbow joint stiffness in healthy subjects. Arch Phys Med Rehabil. 2005;86:82-85. 11. Oatis CA. The use of a mechanical model to describe the stiffness and damping characteristics of the knee joint in healthy adults. Phys Ther. 1993;73:740-749. 12. Olgiati R, Burgunder JM, Mumenthaler M. Increased energy cost of walking in multiple sclerosis: effect of spasticity, ataxia, and weakness. Arch Phys Med Rehabil. 1988;69:846-849. 13. OReilly SC, Jones A, Muir KR, Doherty M. Quadriceps weakness in knee osteoarthritis: the effect on pain and disability. Ann Rheum Dis. 1998;57:588-594. 14. Stein RB, Zehr EP, Lebiedowska MK, Popovic DB, Scheiner A, Chizeck HJ. Estimating mechanical parameters of leg segments in individuals with and without physical disabilities. IEEE Trans Rehabil Eng . 1996;4:201-211. 15. Wartenberg R. Pendulousness of the legs as a diagnostic test. Neurology. 1951;1:18-24. 16. Wright V, Johns RJ. Physical factors concerned with the stiffness of normal and diseased joints. Bull Johns Hopkins Hosp. 1960;106:215-231.

940

J Orthop Sports Phys Ther Volume 36 Number 12 December 2006

Appendix
The stiffness coefficient (k) was calculated from the following: k = IP* n2

where IP* is the moment of inertia of the leg-foot segment with respect to the knee joint and n is the natural frequency of the oscillation.

The damping coefficient (c) is calculated from the following: c = 2 n2 IP*

where is the viscous damping factor found from:

= / ((2)2 + 2)1/2

RESEARCH

and is the logarithmic decrement of the decay:

= 1 ln 1 j j
where is the peak angle of flexion of the jth oscillation.

REPORT

J Orthop Sports Phys Ther Volume 36 Number 12 December 2006

941