JBJS | Article

The Journal of Bone & Joint Surgery, Volume 82, Issue 11

Articles | November 01, 2000

Comparison of Proprioception in Arthritic and Age-Matched Normal Knees*

Lisa M. Koralewicz, M.P.H.†; Gerard A. Engh, M.D.†

Investigation performed at Anderson Orthopaedic Research Center, Alexandria, Virginia
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Anderson Orthopaedic Research Center, P.O. Box 7088, Alexandria, Virginia 22307.

The Journal of Bone & Joint Surgery. 2000; 82:1582-1582

5 Recommendations (Recommend) | 3 Comments | Saved by 3 Users Save Case

text A A A

Abstract

Background: Proprioception - one's ability to sense joint position and joint motion - is affected by factors such as age, muscle fatigue, and osteoarthritis. Most proprioception studies have focused on young active subjects or on recipients of total knee replacements. Few have involved a population with arthritic knees prior to total knee replacement or persons similar in age to patients with advanced knee arthritis who are to have total knee arthroplasty. The purpose of the present study was to determine (1) if proprioception in arthritic knees differs from proprioception in nonarthritic, age-matched, normal knees; (2) if, when proprioception in one knee is reduced by the presence of advanced gonarthrosis, it also is reduced in the contralateral knee irrespective of the presence of arthritis; and (3) if a person's grade of arthritis is associated with his or her level of proprioception.

Methods: This study compared the proprioception levels of a group of 117 patients who were scheduled for total knee arthroplasty because of severe arthritis (mean age, 67.9 years) with those of a control group of forty patients who were recruited from a hospital-based cardiac rehabilitation program and did not have knee arthritis (mean age, 68.3 years). We used a customized Biodex System 2 Multi-Joint Testing and Rehabilitation System to compare proprioception (the threshold to detection of passive motion) between the two groups.

Results: Middle-aged and elderly persons with advanced knee arthritis were significantly less able to detect passive motion of the knee than were middle-aged and elderly persons without knee arthritis. Patients who had arthritis in only one knee had a reduced ability to detect passive motion of both knees. There was no significant association between the radiographic severity of arthritis and the threshold to detection of passive motion in patients with advanced knee arthritis.

Conclusions: Knee proprioception in middle-aged and elderly persons with advanced knee arthritis is reduced in comparison with that in middle-aged and elderly persons without arthritis. Such loss of proprioception is independent of the severity of knee arthritis and may foretell the development of arthritis. When a patient has reduced proprioception with regard to one knee affected by arthritis, he or she also has reduced proprioception with regard to the contralateral knee, independent of the presence or severity of degenerative arthritis. When an investigator is evaluating changes in proprioception after knee arthroplasty, it is best to compare the knee with the patient's untreated knee rather than with age-matched controls.

Figures in this Article

Introduction

Introduction | Materials and Methods | Results | Discussion | References

Knee ligaments in conjunction with the joint capsule provide both a mechanical restraining function and sensory feedback. This sensory feedback controls the dynamic component of joint stability provided by muscle reflex as well as joint position sensibility (knowing where one's joint is in space).

Proprioception encompasses the senses of joint position and joint motion. These senses originate from the stimulation of specialized nerve-endings, or mechanoreceptors, in the joint capsule and ligaments. The receptors convert the mechanical energy of physical deformation into the electrical energy of a nerve action potential⁴. Responding to constant or slow pressures, mechanoreceptors give steady-state information on joint position and also sense motion and angle of rotation⁴.

To assess proprioception, researchers have tested the sense of both joint position and joint motion, or kinesthesia. To test the sense of joint position, investigators have placed subjects' legs in various predetermined angles of flexion; the subjects have then reproduced their perception of the angle of flexion on a visual analog scale^5,8,11,17,18. Researchers also have assessed joint position sense by having subjects reposition their legs in a remembered angle of flexion^2,7,18. To test the sense of joint motion, or kinesthesia, researchers have examined the point at which patients can detect slow passive motion. Measured in degrees of angular displacement, this point (at which the subject detects movement or a change in position) is called the threshold to detection of passive motion^2,6,15,18. The measurement device is usually a dynamometer that slowly, constantly, and passively flexes or extends a person's limb at speeds of no more than 1 degree per second. The threshold to detection of passive motion provides a more objective measure of proprioception than a reproduction on a visual analog scale does.

Proprioception declines with age^10,17, muscle fatigue¹⁸, and articular disease such as osteoarthritis². However, the cause of reduced proprioception remains unclear, as does the temporal sequence of degenerative arthritis and reduced proprioception - that is, it is unknown whether reduced proprioception causes degenerative arthritis because of reduced muscle reflex or if reduced proprioception results from degenerative arthritis and the loss of ligament tension as the joint space collapses.

Most proprioception studies have focused on young active subjects^1,3,7,17,18 or on recipients of total knee replacements^2,6,15,16. Few have involved a population with arthritic knees prior to total knee replacement or persons similar in age to the average arthritic patient. To our knowledge, researchers have not yet addressed whether an elderly patient with advanced knee arthritis has a different kinesthetic capacity than a similarly aged person without gonarthrosis, and they have not investigated whether persons with severely arthritic knees have less kinesthetic sensitivity (larger thresholds to detection of passive motion) than persons with less severely arthritic knees. The purpose of our study was to determine (1) if arthritic knees differ from nonarthritic, age-matched, asymptomatic knees with respect to proprioception; (2) if, when one knee has advanced arthritis, the contralateral knee has altered proprioception, regardless of the presence or absence of gonarthrosis; and (3) if a person's grade of arthritis is associated with his or her level of proprioception.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 1:: Photograph showing the customized test apparatus, which initiates motion at 0.5 degree per second and records the angular displacement when the subject perceives motion or a change in the position of the knee. The subject must then identify the direction of movement.

View Larger

Table I: Group Demographics

*The values are given as the mean, with the range in parentheses.

	Control Group* (N = 40)	Arthritic Group* (N = 117)	P Value
Age (yrs.)	68.3 (43.8-79.8)	67.9 (44.6-91.2)	0.894
Weight (kg)	82.2 (57.6-123.4)	84.7 (49.0-145.1)	0.423
Height (cm)	176.0 (154.9-193.0)	169.4 (149.9-193.0)	0.001
Body-mass index (kg/m²)	26.9 (19.7-38.3)	29.4 (19.5-49.8)	0.003

Add to My JBJS

Table I: Group Demographics

*The values are given as the mean, with the range in parentheses.

	Control Group* (N = 40)	Arthritic Group* (N = 117)	P Value
Age (yrs.)	68.3 (43.8-79.8)	67.9 (44.6-91.2)	0.894
Weight (kg)	82.2 (57.6-123.4)	84.7 (49.0-145.1)	0.423
Height (cm)	176.0 (154.9-193.0)	169.4 (149.9-193.0)	0.001
Body-mass index (kg/m²)	26.9 (19.7-38.3)	29.4 (19.5-49.8)	0.003

View Larger

Table II: Comparisons of Thresholds to Detection of Passive Motion Between Study Groups

*The values are given as the mean, with the range in parentheses.†Compared with the control group.

	Control Group* (N = 40)	Arthritic Group
Side Scheduled for Op.* (N = 117)	P Value†	Side Not Scheduled for Op.* (N = 93)	P Value
Threshold of detection (degrees)
Extension	1.42 (0.26-4.08)	2.57 (0-23)	0.003	2.25 (0-20.41)	0.054
Flexion	1.64 (0.61-4.43)	2.71 (0-16.3)	0.003	2.74 (0.23-17.95)	0.003

Add to My JBJS

Table II: Comparisons of Thresholds to Detection of Passive Motion Between Study Groups

*The values are given as the mean, with the range in parentheses.†Compared with the control group.

	Control Group* (N = 40)	Arthritic Group
Side Scheduled for Op.* (N = 117)	P Value†	Side Not Scheduled for Op.* (N = 93)	P Value
Threshold of detection (degrees)
Extension	1.42 (0.26-4.08)	2.57 (0-23)	0.003	2.25 (0-20.41)	0.054
Flexion	1.64 (0.61-4.43)	2.71 (0-16.3)	0.003	2.74 (0.23-17.95)	0.003

View Larger

Table III: Comparisons of Thresholds to Detection of Passive Motion Within Groups

*The values are given as the mean, with the range in parentheses.

		Control Group	Arthritic Group
Left Side* (N = 40)	Right Side* (N = 40)	P Value	Side Scheduled for Op.* (N = 117)	Side Not Scheduled for Op.* (N = 93)	P Value
Threshold of detection (degrees)
Extension	1.36 (0.28-3.75)	1.48 (0.14-5.15)	0.962	2.57 (0-23)	2.25 (0-20.41)	0.129
Flexion	1.51 (0-3.5)	1.76 (0.5-5.6)	0.357	2.71 (0-16.3)	2.74 (0.23-17.95)	0.357

Add to My JBJS

Table III: Comparisons of Thresholds to Detection of Passive Motion Within Groups

*The values are given as the mean, with the range in parentheses.

		Control Group	Arthritic Group
Left Side* (N = 40)	Right Side* (N = 40)	P Value	Side Scheduled for Op.* (N = 117)	Side Not Scheduled for Op.* (N = 93)	P Value
Threshold of detection (degrees)
Extension	1.36 (0.28-3.75)	1.48 (0.14-5.15)	0.962	2.57 (0-23)	2.25 (0-20.41)	0.129
Flexion	1.51 (0-3.5)	1.76 (0.5-5.6)	0.357	2.71 (0-16.3)	2.74 (0.23-17.95)	0.357

View Larger

Table IV: Gender Differences in Threshold to Detection of Passive Motion Between Groups

*The values for threshold of detection are given as the mean, with the range in parentheses.

	Male*	Female*	P Value
Control group
No. of knees	34	6
Threshold of detection (degrees)
Extension	1.39 (0.26-4.08)	1.57 (0.51-2.8)	0.517
Flexion	1.61 (0.61-4.43)	1.79 (0.64-4.14)	0.698
Arthritic group
Side not scheduled for op.
No. of knees	36	57
Threshold of detection (degrees)
Extension	1.97 (0.09-20.41)	2.44 (0-17.39)	0.074
Flexion	2.31 (0.62-9.83)	3.02 (0.23-17.95)	0.221
Side scheduled for op.
No. of knees	51	66
Threshold of detection (degrees)
Extension	2.24 (0.19-17.03)	2.83 (0-23)	0.029
Flexion	1.64 (0.61-4.43)	2.71 (0-16.3)	0.003

Add to My JBJS

Table IV: Gender Differences in Threshold to Detection of Passive Motion Between Groups

*The values for threshold of detection are given as the mean, with the range in parentheses.

	Male*	Female*	P Value
Control group
No. of knees	34	6
Threshold of detection (degrees)
Extension	1.39 (0.26-4.08)	1.57 (0.51-2.8)	0.517
Flexion	1.61 (0.61-4.43)	1.79 (0.64-4.14)	0.698
Arthritic group
Side not scheduled for op.
No. of knees	36	57
Threshold of detection (degrees)
Extension	1.97 (0.09-20.41)	2.44 (0-17.39)	0.074
Flexion	2.31 (0.62-9.83)	3.02 (0.23-17.95)	0.221
Side scheduled for op.
No. of knees	51	66
Threshold of detection (degrees)
Extension	2.24 (0.19-17.03)	2.83 (0-23)	0.029
Flexion	1.64 (0.61-4.43)	2.71 (0-16.3)	0.003

View Larger

Table V: Thresholds to Detection of Passive Motion According to the Severity of Arthritis in Knees Not Scheduled for an Operation in the Arthritic Group

*The values are given as the mean, with the 95 percent confidence interval in parentheses.†According to one-way analysis of variance.

Severity of Arthritis¹⁴	No. of Knees	Threshold of Detection* (degrees)	P Value†
Extension
Grade 0	18	1.66 (0.93-2.4)	0.419
Grade 1	21	1.67 (1.1-2.2)
Grade 2	15	1.46 (0.90-2.0)
Grade 3	21	2.67 (0.76-4.6)
Flexion
Grade 0	18	3.08 (1.1-5.0)	0.507
Grade 1	21	2.48 (1.5-3.4)
Grade 2	15	1.77 (1.3-2.2)
Grade 3	21	2.60 (1.7-3.5)

Add to My JBJS

Table V: Thresholds to Detection of Passive Motion According to the Severity of Arthritis in Knees Not Scheduled for an Operation in the Arthritic Group

*The values are given as the mean, with the 95 percent confidence interval in parentheses.†According to one-way analysis of variance.

Severity of Arthritis¹⁴	No. of Knees	Threshold of Detection* (degrees)	P Value†
Extension
Grade 0	18	1.66 (0.93-2.4)	0.419
Grade 1	21	1.67 (1.1-2.2)
Grade 2	15	1.46 (0.90-2.0)
Grade 3	21	2.67 (0.76-4.6)
Flexion
Grade 0	18	3.08 (1.1-5.0)	0.507
Grade 1	21	2.48 (1.5-3.4)
Grade 2	15	1.77 (1.3-2.2)
Grade 3	21	2.60 (1.7-3.5)

Materials and Methods

Introduction | Materials and Methods | Results | Discussion | References

We performed a case-control study in which the cases (patients scheduled to have total knee arthroplasty) were selected by virtue of having a particular disease (arthritis requiring total knee arthroplasty). The study group included patients with advanced arthritis in one or both knees who were scheduled to have primary total knee replacement between April 1998 and March 1999 and who had not had prior total knee replacement, unicondylar knee replacement, or high tibial osteotomy. Patients who had had prior arthroscopy or meniscectomy were included. The patients were tested within thirty days prior to knee replacement.

The control group was recruited from a hospital-based cardiac rehabilitation program. These enrollees did not have knee pain or previous knee surgery and were not limited in their regular participation in cardiac rehabilitation activities. Patients were excluded from both the study group and the control group if they had a neuropathic disease, such as insulin-dependent diabetes mellitus, or an autoimmune disease, such as rheumatoid arthritis, Charcot disease, or lupus erythematosus. We also excluded patients with peripheral vascular disease, functional limitations, or difficulty in understanding the English language.

A power analysis was performed to determine appropriate sample size. We selected 1 degree as the smallest mean difference that would be of clinical importance. A standard deviation of 1.1 degrees within groups was selected by referencing results from a similar test protocol¹⁷. We determined that inclusion of thirty-three persons per group would achieve an alpha of 0.05 and a power of 0.95.

Group Demographics

The study group consisted of 117 arthritic patients (sixty-six women and fifty-one men) who were scheduled for total knee arthroplasty. Twenty-four of these patients were scheduled to have a total knee arthroplasty on the contralateral side as well; the other ninety-three patients were not scheduled to have a total knee arthroplasty on the contralateral side. The control group consisted of forty patients (six women and thirty-four men) who met the selection criteria (Table I). The size of the control group was determined by our power analysis for sample size and the number of volunteers from the cardiac rehabilitation program who met the selection criteria. The predominance of men in the control group was due to the relatively small number of women in the cardiac rehabilitation program; it did not reflect a discrepancy in testing effort or a difference between men and women in willingness to participate. The groups were well matched for age (mean, 67.9 years in the study group and 68.3 years in the control group) and weight (mean, 186.7 pounds [84.7 kilograms] in the study group and 181.2 pounds [82.2 kilograms] in the control group). All tests were performed between April 1998 and May 1999.

Measurement of Proprioception of the Knee

Researchers measured proprioception of the knee using a Biodex System 2 Multi-Joint Testing and Rehabilitation System (Biodex Medical Systems, Shirley, New York). Custom software enabled the Biodex to extend or flex the attached knee passively at 0.5 degree per second, a speed compatible with the slowly adapting mechanoreceptors of the knee capsule⁴.

Each subject performed one practice test to become familiar with the test process before completing the trials. A total of twelve randomized trials were completed according to the following protocol. The test subjects were blindfolded and wore headphones to eliminate visual and auditory stimuli from the testing procedure apparatus. They wore shorts to negate any extraneous skin sensation from clothing touching the knee area. Each person sat in the Biodex with the feet in inflatable pressure boots. These boots eliminated any sensation cues from skin or ankle position by immobilizing the foot. With the subject's knee in 45 degrees of flexion, the researcher hooked the boots to the lever arm of the Biodex (Fig. 1). The researcher verified the 45-degree angle of flexion with a goniometer and made fine adjustments to obtain this angle.

The Biodex machine extended or flexed the knee at 0.5 degree per second until the subject detected passive motion or a change in joint position and stopped the motion with a handheld stop button; he or she was then asked to identify the direction (flexion or extension) of the knee movement. The direction of the trial was randomized, and researchers recorded both the stop position (threshold to detection of passive motion) and the direction (measured in degrees of angular displacement) for each trial. Three randomized tests were conducted on one leg and then on the other leg. This was repeated until each patient completed a total of twelve tests: three for flexion of the right knee, three for extension of the right knee, three for flexion of the left knee, and three for extension of the left knee.

For statistical analysis, we grouped the measurements of threshold to detection of passive motion for arthritic knees undergoing total knee replacement (117 knees). We graded seventy-five of the 117 contralateral knees of the arthritic patients with regard to the severity of arthritis as assessed on preoperative anteroposterior and lateral radiographs. The other forty-two contralateral knees were excluded from this analysis because they were scheduled for a total knee arthroplasty (twenty-four knees) or because preoperative radiographs were not available (eighteen knees). We used the technique described by Resnick and Niwayama¹⁴. This technique defines severity on a scale of 0 to 3, with grade 0 representing no radiographic changes of arthritis and grade 3 representing marked joint-space narrowing, osseous collapse, sclerosis, intra-articular osseous bodies, and marked deformity. In the control group, the mean threshold to detection of passive flexion was calculated by adding the flexion on the left and the flexion on the right and dividing by two. The same was done for the extension values.

Large measurements of threshold to detection of passive motion indicated less sensitivity to passive motion than small measurements did. Using the Mann-Whitney test, we compared measurements of threshold to detection of passive motion between the arthritic patients and the control group. Comparisons between the knees of the individual subjects (within-subject comparisons) were performed with use of the Wilcoxon test. The chi-square test was used to discern differences between categorical variables. Data were analyzed with use of SPSS statistical software (version 8.0; SPSS, Chicago, Illinois).

Results

Introduction | Materials and Methods | Results | Discussion | References

The arthritic patients who were to undergo total knee arthroplasty had a significantly larger mean threshold to detection of passive knee motion in both flexion and extension than did the patients without arthritis (the control group) (p < 0.05) (Table II); in other words, the patients without known gonarthrosis had significantly better proprioception. The control group patients exhibited no significant difference in proprioception between the right and left knees (p > 0.05) (Table III). Likewise, the study group, in which the patients had gonarthrosis in one or both knees, did not have significantly different mean scores for threshold to detection of passive motion for their two knees (Table III). Thus, the contralateral knees of patients with advanced gonarthrosis were more similar to the arthritic knees of those patients than to the nonarthritic knees of age-matched patients in the control group.

There were no differences between the control group and the arthritic group with respect to age or weight (Table I). However, the groups differed in gender distribution (Pearson chi-square, p = 0.001). The control group had fewer women: 15 percent of the patients in the control group were women compared with 56 percent in the arthritic group. As stated earlier, this difference in gender representation was related to the relatively small number of women enrolled in the cardiac rehabilitation program. Because the difference was significant, the gender variable could have contributed to or been responsible for the differences in proprioception, but we found no significant differences within the control group between men and women with regard to proprioception in either flexion or extension (Table IV). Because of the small number of women in the control group, we also evaluated the knees not scheduled for an operation in the arthritic group, to look for gender-related differences. There was no difference between the thirty-six knees in men and the fifty-seven knees in women with regard to proprioception in either extension or flexion. Therefore, we concluded that the difference in gender distribution between the two groups was not responsible for the differences in the threshold to detection of passive motion between the groups.

However, when we compared the knees that were scheduled for surgery in the arthritic group, we found that the women (sixty-six knees) had a significantly higher mean threshold to detection of passive motion than did the men (fifty-one knees) in both extension (p = 0.029) and flexion (p = 0.003), indicating that the women were less sensitive to motion. Perhaps this represents a difference in the severity of the disease between men and women. We acknowledge that the area of gender difference warrants further research.

There also was a significant difference between the arthritic and control groups with regard to height (p = 0.001) and body-mass index (p = 0.003). It is likely that the demographic difference in gender between the groups contributed to the difference in height between the groups. We attribute the lower mean body-mass index of our control group to the participants' extended involvement in a rehabilitation exercise and weight-control program.

Additionally, we compared the scores for threshold to detection of passive motion within groups to determine if the direction of the test (flexion or extension) influenced the magnitude of the score. The knees of the patients in the control group and the knees not scheduled for an operation in the study group were more sensitive to an extension arc (p < 0.05) than they were to a flexion arc. The arthritic knees of the patients in the study group differed with regard to the thresholds in extension and flexion, but the difference was not significant.

To determine if there was a linear association between the grade of arthritis and proprioception, we compared the grade of arthritis determined on the preoperative anteroposterior and lateral radiographs of the contralateral knees of the arthritic patients with the scores for the threshold to detection of passive motion. One-way analysis of variance revealed no association between the four grades of arthritis and proprioception (p > 0.05) (Table V). Thus, the loss of proprioception in the knees that were not scheduled for an operation in the patients with gonarthrosis was independent of the radiographic severity of the arthritis in those knees.

Discussion

Introduction | Materials and Methods | Results | Discussion | References

A host of proprioception studies^{2,5,6,9,16,17,19,21} have been performed on patients following total knee arthroplasty in an effort to determine (1) if proprioception returns or further deteriorates following total knee arthroplasty, (2) if the type of implant influences proprioception, and (3) if posterior cruciate retention is beneficial with regard to proprioception. Different investigators obtained different results, but most found no significant difference in proprioception related to cruciate retention and minimal, if any, improvement in proprioception after knee replacement. Likewise, proprioception was minimally influenced by the type of prosthesis (linked compared with unconstrained)^5,21. However, none of these studies included preoperative measurements, raising questions about the validity of their findings.

In addition, most proprioception studies have focused on young active subjects^1,3,7,17,18 or on recipients of total knee replacements^2,6,15,16. To our knowledge, our study was the first to compare the threshold to detection of passive motion of middle-aged to elderly patients with arthritis with the threshold in an age-matched population without arthritis (mean ages, 67.9 and 68.3 years, respectively). It also differed from previous studies in that proprioception was measured preoperatively before the confounding influences of surgery.

In answer to our original questions, we found that proprioception in arthritic knees differed from proprioception in nonarthritic, age-matched, normal knees. Despite the subjects being well matched in age, they differed significantly in terms of the scores for the threshold to detection of passive motion (Table II), with the control patients having significantly better proprioception. These findings were in agreement with those of Barrett et al.⁵, who reported that osteoarthritic subjects were significantly less accurate in reproducing joint position angles with a visual analog model (the alternative method for measuring proprioception) than were subjects with normal knees.

In addition, our results showed that, when proprioception was reduced in one knee that had advanced gonarthrosis, it also was reduced in the contralateral knee, irrespective of the presence of arthritis in that knee (Table III). The results also showed that the grade, or severity, of arthritis in the contralateral knee of an arthritic patient was not linearly associated with his or her level of proprioception (Table V). These findings suggest that loss of proprioception might occur prior to the development of the structural changes of arthritis that can be identified on routine anteroposterior radiographs.

The findings of studies of proprioception after total knee arthroplasty substantiate our observations. Barrack et al.² compared joint motion sensibilities (kinesthesia) of patients who had had a total knee replacement with those in a small group of age-matched controls. The authors found that, compared with the age-matched control group, the patients who had undergone total knee arthroplasty had less sensitive kinesthetic abilities in both the treated and the untreated knees. Barrack et al. also demonstrated that, following total knee arthroplasty, the treated and untreated knees did not differ with regard to the threshold to detection of passive motion. Although the authors stated that all patients had articular disease in the untreated knee, they did not qualify the extent of the arthritis. Barrack et al. obtained no preoperative proprioception measurements, and the alterations of surgery could have confounded their results.

Our results raise additional questions about the factors that underlie the loss of proprioception. The confounding variables associated with loss of proprioception and arthritis make it difficult to determine the impact of any single variable on the overall reduction of proprioception associated with degenerative arthritis. With degenerative arthritis comes atrophy of neural elements^13,15, loss of tension on collateral and capsular ligaments, altered loading patterns across the knee, and muscle atrophy from decreased patient activities. Each of these factors, independently or in combination, could lead to loss of proprioception. Schultz et al.¹⁵ reported that Golgi-tendon-like mechanoreceptors identifiable in autopsy and amputation knee specimens were almost nonexistent in the presence of advanced arthritis. However, our data clearly demonstrated that patients with advanced degenerative arthritis of one knee also lost proprioception in the contralateral knee, with or without radiographic evidence of osteoarthritis or structural changes associated with arthritis in that knee. This led us to speculate that the physiological changes associated with degenerative arthritis, such as muscle atrophy and loss of ligament tension, could not by themselves explain the loss of proprioception.

Our study also showed that proprioception was markedly better in persons of similar age who did not have degenerative arthritis; thus, aging alone could not account for the reduction in proprioception. Perhaps the age-related reduction in detection of passive motion is related to a reduction in the activity level of some individuals with age^5,18. There is a loss of sensory receptors in the cruciate ligaments of patients with advanced arthritis of the knee^13,15; these mechanoreceptors could be lost secondary to atrophy from a reduction in activities due to the arthritis of the knee or from enzymatic damage associated with cartilage degradation. A loss of mechanoreceptors could explain the development of degenerative arthritis in much the same way as sensory deprivation leads to destructive arthritis with Charcot joints. There could be a threshold of sensory recognition for a person's ability to protect the knee joint through neuromotor reflexes, and, when exceeded, the process of joint degeneration could be initiated. The temporal relationship between proprioception and gonarthrosis requires further investigation.

There also is some debate in the literature regarding the origin of the mechanoreceptors that are primarily responsible for proprioception about the human knee⁴. Barrack et al.¹ investigated the mechanoreceptors within the capsule by examining the gait of patients after their knees had been anesthetized with intra-articular injection of 2 percent lidocaine. They compared these knees with ones in which saline solution had been injected and found that functional ability was not altered by anesthesia of the joint capsule. However, mechanoreceptors within the cruciate ligaments may not have been anesthetized since they are intra-articular but extrasynovial. Some investigators have contended that mechanoreceptors responsible for proprioception arise in the skin⁴, whereas others have proposed that they are within the muscle¹¹ or the ligaments^12,14. Mechanoreceptors could be present in all of these structures. To address this issue, we isolated the knee joint as much as possible, so that the proprioceptive input came predominantly from the ligaments and capsular structures of the knee. We did this by minimizing the neurological input from the skin. The inflatable pressure boots worn by the test subjects prevented skin stretch deformation of the ankle joint; the shorts worn by the subjects negated extraneous sensations from movement of clothing about the knee.

Since there was no significant difference in proprioception between the contralateral and diseased knees of the arthritic patients, it appears that the contralateral knees of patients undergoing unilateral total knee arthroplasty can be used as a valid control for measuring changes in proprioception resulting from total knee arthroplasty. Age-matched controls have different proprioception than subjects with arthritis and thus should not be used to evaluate changes in proprioception due to total knee arthroplasty. The best information can be obtained when preoperative proprioception studies with use of the contralateral knee as a control are followed by repeat studies after full recovery from total knee arthroplasty.

In summary, we found that the proprioceptive ability of patients with advanced degenerative arthritis of one or both knees differs from that of age-matched controls and that such patients have reduced proprioception in the contralateral knee, whether or not it has radiographic evidence of osteoarthritis. These results confirm the strong association between reduced proprioception and osteoarthritis but leave questions regarding the temporal relationship unanswered. Additional studies are needed to determine if proprioception can be improved by restoration of physical activity.

References

Introduction | Materials and Methods | Results | Discussion | References

Barrack, R. L.; Skinner, H. B.; Brunet, M. E.; and Haddad, R. J., Jr.: Functional performance of the knee after intraarticular anesthesia. Am. J. Sports Med.,11: 258-261, 1983.11258 1983 [PubMed]

Barrack, R. L.; Skinner, H. B.; Cook, S. D.; and Haddad, R. J., Jr.: Effect of articular disease and total knee arthroplasty on knee joint-position sense. J. Neurophysiol.,50: 684-687, 1983.50684 1983 [PubMed]

Barrack, R. L.; Skinner, H. B.; Brunet, M. E.; and Cook, S. D.: Joint kinesthesia in the highly trained knee. J. Sports Med. and Phys. Fit.,24: 18-20, 1984.2418 1984

Barrack, R. L., and Skinner, H. B.: The sensory function of knee ligaments. In Knee Ligaments: Structure, Function, Injury, and Repair, pp. 95-114. Edited by D. Daniel, W. H. Akeson, and J. J. O'Connor. New York, Raven Press, 1990.

Barrett, D. S.; Cobb, A. G.; and Bentley, G.: Joint proprioception in normal, osteoarthritic and replaced knees. J. Bone and Joint Surg.,73-B(1): 53-56, 1991.73-B(1)53 1991

Cash, R. M.; Gonzalez, M. H.; Garst, J.; Barmada, R.; and Stern, S. H.: Proprioception after arthroplasty: role of the posterior cruciate ligament. Clin. Orthop.,331: 172-178, 1996.331172 1996 [PubMed]

Goodwin, G. M.; McCloskey, D. I.; and Matthews, P. B.: The persistence of appreciable kinesthesia after paralyzing joint afferents but preserving muscle afferents. Brain Res.,37: 326-329, 1972.37326 1972 [PubMed]

Horch, K. W.; Clark, F. J.; and Burgess, P. R.: Awareness of knee joint angle under static conditions. J. Neurophysiol.,38: 1436-1447, 1975.381436 1975 [PubMed]

Ishii, Y.; Terajima, K.; Terashima, S.; Bechtold, J. E.; and Laskin, R. S.: Comparison of joint position sense after total knee arthroplasty. J. Arthroplasty,12: 541-545, 1997.12541 1997 [PubMed]

Jacobsen, K.: Osteoarthritis following insufficiency of the cruciate ligaments in man. A clinical study. Acta Orthop. Scandinavica,48: 520-526, 1977.48520 1977

Kaplan, F. S.; Nixon, J. E.; Reitz, M.; Rindfleisch, L.; and Tucker, J.: Age-related changes in joint proprioception and sensation of joint position. Acta Orthop. Scandinavica,56: 72-74, 1985.5672 1985

Kennedy, J. C.; Alexander, I. J.; and Hayes, K. C.: Nerve supply of the human knee and its functional importance. Am. J. Sports Med.,10: 329-335, 1982.10329 1982 [PubMed]

Kleinbart, F. A.; Bryk, E.; Evangelista, J.; Scott, W. N.; and Vigorita, V. J.: Histologic comparison of posterior cruciate ligaments from arthritic and age-matched knee specimens. J. Arthroplasty,11: 726-731, 1996.11726 1996 [PubMed]

Resnick, D., and Niwayama, G.: Diagnosis of Bone and Joint Disorders, with Emphasis on Articular Abnormalities. Philadelphia, W. B. Saunders, 1981.

Schultz, R. A.; Miller, D. C.; Kerr, C. S.; and Micheli, L.: Mechanoreceptors in human cruciate ligaments. A histological study. J. Bone and Joint Surg.,66-A: 1072-1076, Sept 1984.66-A1072 1984

Shoji, H.; Wolf, A.; Packard, S.; and Yoshino, S.: Cruciate retained and excised total knee arthroplasty. A comparative study in patients with bilateral total knee arthroplasty. Clin. Orthop.,305: 218-222, 1994.305218 1994 [PubMed]

Simmons, S.; Lephart, S.; Rubash, H.; Borsa, P.; and Barrack, R. L.: Proprioception following total knee arthroplasty with and without the posterior cruciate ligament. J. Arthroplasty,11: 763-768, 1996.11763 1996 [PubMed]

Skinner, H. B.; Barrack, R. L.; and Cook, S. D.: Age related decline in proprioception. Clin. Orthop.,184: 208-211, 1984.184208 1984 [PubMed]

Skinner, H. B.; Barrack, R. L.; Cook, S. D.; and Haddad, R. J., Jr.: Joint position sense in total knee arthroplasty. J. Orthop. Res.,1: 276-283, 1984.1276 1984 [PubMed]

Skinner, H. B.; Wyatt, M. P.; Hodgdon, J. A.; Conard, D. W.; and Barrack, R. L.: Effect of fatigue on joint position sense of the knee. J. Orthop. Res.,4: 112-118, 1986.4112 1986 [PubMed]

Warren, P.; Olanlokun, T.; Cobb, A. G.; and Bentley, G.: Proprioception after knee arthroplasty. The influence of prosthetic design. Clin. Orthop.,297: 182-187, 1993.297182 1993 [PubMed]

View Large | Download Slide(.ppt)
Add to My JBJS

View Larger

Table I: Group Demographics

*The values are given as the mean, with the range in parentheses.

	Control Group* (N = 40)	Arthritic Group* (N = 117)	P Value
Age (yrs.)	68.3 (43.8-79.8)	67.9 (44.6-91.2)	0.894
Weight (kg)	82.2 (57.6-123.4)	84.7 (49.0-145.1)	0.423
Height (cm)	176.0 (154.9-193.0)	169.4 (149.9-193.0)	0.001
Body-mass index (kg/m²)	26.9 (19.7-38.3)	29.4 (19.5-49.8)	0.003

Add to My JBJS

Table I: Group Demographics

*The values are given as the mean, with the range in parentheses.

	Control Group* (N = 40)	Arthritic Group* (N = 117)	P Value
Age (yrs.)	68.3 (43.8-79.8)	67.9 (44.6-91.2)	0.894
Weight (kg)	82.2 (57.6-123.4)	84.7 (49.0-145.1)	0.423
Height (cm)	176.0 (154.9-193.0)	169.4 (149.9-193.0)	0.001
Body-mass index (kg/m²)	26.9 (19.7-38.3)	29.4 (19.5-49.8)	0.003

View Larger

Table II: Comparisons of Thresholds to Detection of Passive Motion Between Study Groups

*The values are given as the mean, with the range in parentheses.†Compared with the control group.

	Control Group* (N = 40)	Arthritic Group
Side Scheduled for Op.* (N = 117)	P Value†	Side Not Scheduled for Op.* (N = 93)	P Value
Threshold of detection (degrees)
Extension	1.42 (0.26-4.08)	2.57 (0-23)	0.003	2.25 (0-20.41)	0.054
Flexion	1.64 (0.61-4.43)	2.71 (0-16.3)	0.003	2.74 (0.23-17.95)	0.003

Add to My JBJS

Table II: Comparisons of Thresholds to Detection of Passive Motion Between Study Groups

*The values are given as the mean, with the range in parentheses.†Compared with the control group.

	Control Group* (N = 40)	Arthritic Group
Side Scheduled for Op.* (N = 117)	P Value†	Side Not Scheduled for Op.* (N = 93)	P Value
Threshold of detection (degrees)
Extension	1.42 (0.26-4.08)	2.57 (0-23)	0.003	2.25 (0-20.41)	0.054
Flexion	1.64 (0.61-4.43)	2.71 (0-16.3)	0.003	2.74 (0.23-17.95)	0.003

View Larger

Table III: Comparisons of Thresholds to Detection of Passive Motion Within Groups

*The values are given as the mean, with the range in parentheses.

		Control Group	Arthritic Group
Left Side* (N = 40)	Right Side* (N = 40)	P Value	Side Scheduled for Op.* (N = 117)	Side Not Scheduled for Op.* (N = 93)	P Value
Threshold of detection (degrees)
Extension	1.36 (0.28-3.75)	1.48 (0.14-5.15)	0.962	2.57 (0-23)	2.25 (0-20.41)	0.129
Flexion	1.51 (0-3.5)	1.76 (0.5-5.6)	0.357	2.71 (0-16.3)	2.74 (0.23-17.95)	0.357

Add to My JBJS

Table III: Comparisons of Thresholds to Detection of Passive Motion Within Groups

*The values are given as the mean, with the range in parentheses.

		Control Group	Arthritic Group
Left Side* (N = 40)	Right Side* (N = 40)	P Value	Side Scheduled for Op.* (N = 117)	Side Not Scheduled for Op.* (N = 93)	P Value
Threshold of detection (degrees)
Extension	1.36 (0.28-3.75)	1.48 (0.14-5.15)	0.962	2.57 (0-23)	2.25 (0-20.41)	0.129
Flexion	1.51 (0-3.5)	1.76 (0.5-5.6)	0.357	2.71 (0-16.3)	2.74 (0.23-17.95)	0.357

View Larger

Table IV: Gender Differences in Threshold to Detection of Passive Motion Between Groups

*The values for threshold of detection are given as the mean, with the range in parentheses.

	Male*	Female*	P Value
Control group
No. of knees	34	6
Threshold of detection (degrees)
Extension	1.39 (0.26-4.08)	1.57 (0.51-2.8)	0.517
Flexion	1.61 (0.61-4.43)	1.79 (0.64-4.14)	0.698
Arthritic group
Side not scheduled for op.
No. of knees	36	57
Threshold of detection (degrees)
Extension	1.97 (0.09-20.41)	2.44 (0-17.39)	0.074
Flexion	2.31 (0.62-9.83)	3.02 (0.23-17.95)	0.221
Side scheduled for op.
No. of knees	51	66
Threshold of detection (degrees)
Extension	2.24 (0.19-17.03)	2.83 (0-23)	0.029
Flexion	1.64 (0.61-4.43)	2.71 (0-16.3)	0.003

Add to My JBJS

Table IV: Gender Differences in Threshold to Detection of Passive Motion Between Groups

*The values for threshold of detection are given as the mean, with the range in parentheses.

	Male*	Female*	P Value
Control group
No. of knees	34	6
Threshold of detection (degrees)
Extension	1.39 (0.26-4.08)	1.57 (0.51-2.8)	0.517
Flexion	1.61 (0.61-4.43)	1.79 (0.64-4.14)	0.698
Arthritic group
Side not scheduled for op.
No. of knees	36	57
Threshold of detection (degrees)
Extension	1.97 (0.09-20.41)	2.44 (0-17.39)	0.074
Flexion	2.31 (0.62-9.83)	3.02 (0.23-17.95)	0.221
Side scheduled for op.
No. of knees	51	66
Threshold of detection (degrees)
Extension	2.24 (0.19-17.03)	2.83 (0-23)	0.029
Flexion	1.64 (0.61-4.43)	2.71 (0-16.3)	0.003

View Larger

Table V: Thresholds to Detection of Passive Motion According to the Severity of Arthritis in Knees Not Scheduled for an Operation in the Arthritic Group

*The values are given as the mean, with the 95 percent confidence interval in parentheses.†According to one-way analysis of variance.

Severity of Arthritis¹⁴	No. of Knees	Threshold of Detection* (degrees)	P Value†
Extension
Grade 0	18	1.66 (0.93-2.4)	0.419
Grade 1	21	1.67 (1.1-2.2)
Grade 2	15	1.46 (0.90-2.0)
Grade 3	21	2.67 (0.76-4.6)
Flexion
Grade 0	18	3.08 (1.1-5.0)	0.507
Grade 1	21	2.48 (1.5-3.4)
Grade 2	15	1.77 (1.3-2.2)
Grade 3	21	2.60 (1.7-3.5)

Add to My JBJS

Table V: Thresholds to Detection of Passive Motion According to the Severity of Arthritis in Knees Not Scheduled for an Operation in the Arthritic Group

*The values are given as the mean, with the 95 percent confidence interval in parentheses.†According to one-way analysis of variance.

Severity of Arthritis¹⁴	No. of Knees	Threshold of Detection* (degrees)	P Value†
Extension
Grade 0	18	1.66 (0.93-2.4)	0.419
Grade 1	21	1.67 (1.1-2.2)
Grade 2	15	1.46 (0.90-2.0)
Grade 3	21	2.67 (0.76-4.6)
Flexion
Grade 0	18	3.08 (1.1-5.0)	0.507
Grade 1	21	2.48 (1.5-3.4)
Grade 2	15	1.77 (1.3-2.2)
Grade 3	21	2.60 (1.7-3.5)