Proprioception allows humans to differentiate position and motion of the body, limbs, and joints. At the knee, proprioception is mediated by feedback from specialized receptors in intraarticular and extraarticular musculoskeletal tissues.
Knee proprioception has been studied by measuring the capacity of blinded volunteers to reproduce specific angles of knee flexion1–4,9,10,14,21,22–30 or detect the initiation of knee motion from some fixed flexion angle.2,6,7,15,19–28 At the simplest level, proprioception examinations measure the ability of humans to sense the position of the knee or a change in the position of the joint.
People with unilateral or bilateral osteoarthritis of the knee have poorer knee proprioception than individuals who do not have osteoarthritis of the knees.4,15,17,19 It has yet to be established whether diminished proprioception is a cause or consequence of osteoarthritis of the knees. It also is unclear why proprioception differs among persons with osteoarthritis of the knees and why proprioception differs between the knees of a person with osteoarthritis of the knees.
To better comprehend the factors that influence proprioception in osteoarthritis of the knee, proprioception was studied in patients scheduled for knee arthroplasty. We hypothesized that knees with more severe clinical involvement would have a greater deficit, and that those considered candidates for unicondylar arthroplasty, because of predominantly monocompartmental tibiofemoral disease, would be less affected. This information is needed to provide more clear insight into the importance of proprioception in patient selection, expectations for outcomes, and choice of procedure.
MATERIALS AND METHODS
Between April 1998 and December 1999, 119 patients with osteoarthritis of the knees had bilateral proprioception examinations as part of preadmission testing before knee arthroplasty. Patients who had a prior knee arthroplasty, diabetes, past or present neuropathy, poor command of the English language, or were physically unable to do the test were excluded from the study. The 119 study patients represented 57% of patients having knee arthroplasty for osteoarthritis for the first time under the direction of the senior author (GAE).
The patient was seated on the chair of a customized testing apparatus (System 2 Multi-Joint Testing and Rehabilitation System, Biodex Medical Systems, Shirley, NY).15 The leg being examined was connected to the lever arm of the testing machine via an inflatable pressure boot. Patients wore shorts, headphones, and a blindfold to limit feedback from other senses. Before each examination, the examiner (one of three physical therapists) informed the subject that the leg would be flexed or extended at some random time after a tactile alert cue and was asked that he or she press a handheld stop button once motion was sensed. During each trial, the examiner positioned the knee in a standard flexion angle of 45°, recorded the initial position of the lever arm from the monitor of the computer controller, cued the examinee for alertness, paused, released a switch directing the controller to flex or extend the knee at a rate of 0.5° per second, and recorded the final position of the lever arm from the readout after the subject had pressed the stop button. Three extension trials (decreasing flexion angle) and three flexion trials (increasing flexion angle) were done nonconsecutively with the right and left limbs.
Data collected during the proprioception examinations were processed as follows. For each trial, the difference between the initial and final positions of the lever arm was converted to angular degrees and designated as the threshold to detection of passive motion for that trial. For each knee, a proprioceptive threshold for flexion was computed by averaging the threshold to detection of passive motion measurements obtained during knee flexion, and a proprioceptive threshold for extension was computed by averaging the threshold measurements noted during knee extension.
Demographic data, including age, gender, and body mass index, were compiled with the aid of a clinical database (Table 1). The physical therapist had completed a Knee Society function score13 on the day of preadmission (and proprioception) testing and had also asked each subject to rate themselves as very active, moderately active, inactive, or disabled. Each knee was categorized as either having prior surgery or no prior surgery. Range of motion (ROM) and Knee Society13 AP and mediolateral stability data collected in the preoperative period were reviewed. One observer (ESS) graded the severity of osteoarthritis from weightbearing AP knee radiographs using the scale described by Duwelius and Connolly.8 Deformity was quantified as the absolute difference between 7° valgus and the radiographic tibiofemoral varus-valgus angle. For each patient, the more symptomatic and the less symptomatic of the two knees were determined by patient preference.
One-hundred fifty-seven of the 238 knees were replaced in the year after proprioception examination. One-hundred ten knees (93 patients) had a TKA and 47 knees (37 patients) had a unicondylar arthroplasty. Criteria for implantation of unicondylar arthroplasty included intact cruciate ligaments and degeneration of Outerbridge Grade III or less18 in the unresurfaced tibiofemoral compartment. Twenty (17 patients) of the 110 total knee arthroplasty knees were judged to have an attenuated, deficient, or absent anterior cruciate ligament on intraoperative inspection. Three knees (3 patients) had an attenuated or deficient posterior cruciate ligament.
Multiple linear regression analysis was done with the proprioceptive threshold for flexion or extension as the dependent variable and the following as independent variables: age; gender; body mass index; self-assessed activity level; Knee Society function score; presence or absence of prior surgery; severity of osteoarthritis; tibiofemoral varus-valgus angular deformity; total ROM; AP drawer of greater or less than 5 mm; and mediolateral laxity of greater or less than 6° opening. The multiple regression analyses were done using SPSS (Version 8.0, SPSS Inc, Chicago, IL), first by entering each independent variable into the analysis and then by entering only those variables that were significantly associated with the proprioceptive threshold in the original regression analysis. The paired t test or one-way analysis of variance (ANOVA) was used to compare datasets with normal distributions. The Wilcoxon signed ranks test (for paired data sets) or Mann-Whitney U test (for unpaired data sets) were used to compare data that failed to assume a normal distribution. Probability values of 0.05 or less were considered significant.
For the 238 knees, proprioceptive thresholds averaged 3.0° ± 3.1° (range, 0.1°-27.5°) in flexion and 2.7° ± 3.4° (range, 0.1°–23.5°) in extension. The median differences between observations from the same knee were 0.7° for flexion and 0.6° for extension.
The 11 factors featured in the multiple regression equation explained no more than 12% (R2 = 0.12) of the variations in proprioceptive thresholds for flexion and extension (Table 2). Age was the variable most strongly associated with proprioceptive thresholds, as younger patients generally responded sooner to the flexion or extension stimulus than older patients. Thresholds for flexion also increased with decreasing ROM, indicating slower response times to the flexion stimulus for subjects with decreased motion. Thresholds for extension increased not only with patient age but also with male patients and diminishing activity level.
A comparison of data collected from the right and left knees of each patient showed that proprioception was poorer in the more symptomatic of the two knees (Table 3). The mean proprioceptive thresholds for flexion and extension were higher for the more symptomatic knee than for the less symptomatic knee (p = 0.03 for flexion and p < 0.01 for extension, Wilcoxon signed ranks test). Compared with the less symptomatic knee, the more symptomatic knee had a more severe radiographic osteoarthritis grade (p < 0.01, Wilcoxon signed ranks test), greater angular deformity (p < 0.01, paired t test), smaller ROM (p < 0.01, paired t test), and a higher incidence of prior surgery (p < 0.01, Wilcoxon signed ranks test).
Preoperative thresholds did not differ among the 47 knees that were implanted with a unicondylar arthroplasty and the 110 knees that had a TKA (p = 0.30 for flexion and p = 0.94 for extension, Mann-Whitney U test; Table 4). With respect to the 11 independent variables (age, gender, body mass index, activity level, function score, prior surgery, radiographic osteoarthritis grade, tibiofemoral angular deformity, ROM, AP instability, mediolateral instability) examined in the multiple regression analysis, knees that needed TKAs differed from knees that needed unicondylar arthroplasties only in that they had more severe osteoarthritis (Grade 2.5 ± 0.5 versus Grade 2.2 ± 0.4; p < 0.01, Mann-Whitney U test), they had less motion (109° ± 15° versus 115° ± 13°; p = 0.05, one-way ANOVA with Tamhane’s post hoc test), and the patients had higher body mass indices (31 ± 6 versus 27 ± 4, p < 0.01, one-way ANOVA with Tamhane’s post hoc test). Knees that were judged (intraoperatively) to have an attenuated, deficient, or absent anterior cruciate ligament (ACL) tended to have poorer preoperative proprioception when we reviewed all replaced knees (flexion, 3.2° ± 1.7° versus 3.0° ± 3.6°, p = 0.08; extension, 4.3° ± 5.9° versus 2.7° ± 3.2°, p = 0.17) and knees that had TKAs (flexion, 3.2° ± 1.8° versus 3.1° ± 3.7°, p = 0.18; extension, 4.5° ± 6.0° versus 2.5° ± 3.5°, p = 0.10).
Proprioceptive thresholds tended to be lower in magnitude and less variable in knees that were not replaced, but did not differ significantly compared with thresholds of the knees that had unicondylar arthroplasties (p = 0.32 for flexion and p = 0.18 for extension, Mann-Whitney U test) or knees that had TKAs (p = 0.87 for flexion and p = 0.13 for extension, Mann-Whitney U test). Nonreplaced knees differed from knees that had unicondylar arthroplasties in that they had less deformity (6° ± 4° versus 9° ± 3°; p < 0.01, one-way ANOVA with Tukey’s post hoc test), less severe osteoarthritis (Grade 1.8 ± 0.7 versus Grade 2.2 ± 0.4; p < 0.01, Mann-Whitney U test), and were less likely to have had prior surgery (11% versus 36%; p < 0.01, Mann-Whitney U test). Compared with knees that had TKAs, nonreplaced knees had more motion (118° ± 15° versus 109° ± 15°; p < 0.01, one-way ANOVA with Tukey’s post hoc test), less deformity (6° ± 4° versus 10° ± 5°; p < 0.01, one-way ANOVA with Tukey’s post hoc test), less severe osteoarthritis (Grade 1.8 ± 0.7 versus Grade 2.5 ± 0.5; p < 0.01, Mann-Whitney U test), less AP instability (0% versus 6%; p = 0.05, Mann-Whitney U test), and were less likely to have had prior surgery (11% versus 41%; p < 0.01, Mann-Whitney U test).
Demographic variables examined in our study were age, gender, and body mass index. Others have reported that thresholds to detection of passive motion increase with age in general populations2,19,26,27 and populations with osteoarthritis.23 In our study of patients who typically had moderate to end-stage osteoarthritis in at least one knee, we found that statistically significant, positive correlations of weak strength existed between age and proprioceptive thresholds to detection of passive flexion and extension. Males tended to respond slower to an extension stimulus than did females. Proprioceptive thresholds were not associated with body mass index.
Across broad populations of elderly individuals, activity levels and functional capacities decrease with age. Pai et al19 reported a moderate correlation between proprioceptive thresholds and the physical functional score of the Western Ontario and McMaster University Osteoarthritis Index.5 Although investigators in that study used a different index than we did to measure functional capacity, they also did bivariate correlation analyses that may be swayed by the inherent association between aging and the physical function score. Although we did identify significant correlation between the self-assessed activity level and the proprioceptive threshold for extension, our multivariate regression analysis showed that proprioceptive deterioration was associated more with increasing age than diminishing functional capacity or activity level. This finding may either attest to the difficulty in quantifying activity levels and functional capacities in an accurate and objective fashion, or suggest that changes apart from diminishing activity levels and functional capacities are responsible for the proprioceptive deterioration that accompanies aging in populations with osteoarthritis.
Given the proprioceptive deficits common to knees with osteoarthritis, it would be suspected that proprioception deteriorates as the severity of osteoarthritis increases and that proprioception therefore would be directly or indirectly associated with variables that often accompany osteoarthritis (such as increasing incidence of prior surgical intervention, increasing radiographic joint space narrowing, increasing tibiofemoral deformity, decreasing ROM, and increasing AP or mediolateral joint instability). However, our multivariate analysis consistently failed to link proprioceptive thresholds with clinical variables related to the knee under examination. Only with regard to proprioceptive thresholds for flexion did we find an association between proprioception and a knee variable (proprioception deteriorated slightly as ROM decreased). Pai et al19 did not identify a significant association between proprioceptive thresholds and osteoarthritis severity, joint laxity, or ROM in their bivariate correlation analyses of knees with osteoarthritis. Wada et al29 found that angular reproduction errors before TKA did not correlate with limb alignment or varus-valgus laxity.
Although symptoms specific to the knee did not correlate with proprioceptive thresholds across the entire study group, we did find that in the same patient the more symptomatic of the two knees had significantly poorer proprioception when compared with the less symptomatic knee. Although proprioception would be suspected to be worse in the more troublesome knee, this is the first study of subjects with osteoarthritis of the knees to document a proprioceptive difference between the bilateral knees compared with symptomatology. Sharma et al23 compared bilateral knees in 28 subjects with unilateral osteoarthritis and could not identify a statistical difference between proprioception of the knees with and without arthritis.
Less extensive osteoarthritic degeneration and retention of functionally intact ACLs and PCLs would seem to predispose knees that need unicondylar arthroplasty to have better proprioception before and after arthroplasty in comparison to knees that need TKAs. In a study of proprioceptive thresholds measured after knee arthroplasty, Simmons et al25 reported that 10 knees with unicondylar arthroplasties were not statistically different from 15 knees with posterior cruciate-retaining TKAs or 13 knees with posterior cruciate-sacrificing TKAs. Simmons et al25 did not measure proprioception before arthroplasty. Despite the fact that the knees of patients who had unicondylar and TKAs in our study differed preoperatively and intraoperatively, the knees did not differ when we compared their preoperative thresholds with detection of passive motion. It does not seem unreasonable to compare proprioceptive thresholds measured after TKA with thresholds measured after unicondylar arthroplasty, so long as other factors known to influence thresholds do not differ between the populations. It seems important that the ages of patients having unicondylar arthroplasty parallel those of patients having TKAs if proprioception is to be compared after surgery in the absence of preoperative proprioception data. The resurgence in popularity of unicondylar arthroplasty and the level of contemporary interest in minimally invasive arthroplasty techniques may stimulate increased attention to the impact of knee arthroplasty on proprioception. In contrast to Wada et al,29 who found that proprioception was significantly poorer when the ACL was intact at TKA, we found that proprioception tended to be poorer when the anterior cruciate was deficient.
Our findings show the difficulty one experiences in trying to predict proprioceptive thresholds for a knee based on knowledge of standard data that clinicians normally document before surgery. We found that no more than 12% of the proprioceptive variation that existed between knees could be explained by the variables featured in our regression analysis. The ability to identify patients with osteoarthritis who are more likely to have poorer proprioception may be valuable to the clinician in several ways, such as selecting the type of knee arthroplasty, evaluating the impact of knee arthroplasty and rehabilitation protocols on proprioception, and in further probing the poorly understood connection between osteoarthritis and proprioception at the knee. Attention to parameters that are not typically assessed in the clinical setting (sensorimotor deterioration,12,16,31 muscular atrophy,11,12 pharmaceutical use19) may be required to foster better understanding of proprioception in the osteoarthritic knee.
We thank Fran Preidis, Don Robinson, and Nancy Pugh of the INOVA Mount Vernon Hospital Joint Replacement Center for coordinating and doing the examinations.
1. Attfield SF, Wilton TJ, Pratt DJ, Sambatakakis A: Soft-tissue balance and recovery of proprioception after total knee replacement. J Bone Joint Surg 78B:540–545, 1996.
2. Barrack RL, Skinner HB, Cook SD, Haddad RJ: Effect of articular disease and total knee arthroplasty on knee joint-position sense. J Neurophysiol 50:683–687, 1983.
3. Barrett DS: Proprioception and function after anterior cruciate reconstruction. J Bone Joint Surg 73B:833–837, 1991.
4. Barrett DS, Cobb AG, Bentley G: Joint proprioception in normal, osteoarthritic and replaced knees. J Bone Joint Surg 73B:53–56, 1991.
5. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW: Validation study of the 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 15:1833–1840, 1988.
6. Cash RM, Gonzalez MH, Garst J, Barmada R, Stern SH: Proprioception after arthroplasty: Role of the posterior cruciate ligament. Clin Orthop 331:172–178, 1996.
7. Corrigan JP, Cashman WF, Brady MP: Proprioception in the cruciate deficient knee. J Bone Joint Surg 74B:247–250, 1992.
8. Duwelius PJ, Connolly JF: Closed reduction of tibial plateau fractures: A comparison of functional and roentgenographic end results. Clin Orthop 230:116–126, 1988.
9. Fuchs S, Frisse D, Tibesku CO, Laass H, Rosenbaum D: Proprioceptive function, clinical results, and quality of life after unicondylar sledge prostheses. Am J Phys Med Rehabil 81:478–482, 2002.
10. Fuchs S, Thorwesten L, Niewerth S: Proprioceptive function in knees with and without total knee arthroplasty. Am J Phys Med Rehabil 78:39–45, 1999.
11. Hurley MV: The role of muscle weakness in the pathogenesis of osteoarthritis. Rheum Dis Clin North Am 25:283–298, 1999.
12. Hurley MV, Rees J, Newham DJ: Quadriceps function, proprioceptive acuity and functional performance in healthy young, middle-aged and elderly subjects. Age Ageing 27:55–62, 1998.
13. Insall JN, Dorr LD, Scott RD, Scott WN: Rationale of the Knee Society clinical rating system. Clin Orthop 248:13–14, 1989.
14. Kaplan FS, Nixon JE, Reitz M, Rindfleish L, Tucker J: Age-related changes in proprioception and sensation of joint position. Acta Orthop Scand 56:72–74, 1985.
15. Koralewicz LM, Engh GA: Comparison of proprioception in arthritic and age-matched normal knees. J Bone Joint Surg 82A:1582–1588, 2000.
16. Li SC, Huxhold O, Schmiedek F: Aging and attenuated processing robustness: Evidence from cognitive and sensorimotor functioning. Gerontology 50:28–34, 2004.
17. Marks R, Quinney HA, Wessel J: Proprioceptive sensibility in women with normal and osteoarthritic knee joints. Clin Rheumatol 12:170–175, 1993.
18. Outerbridge RE: The etiology of chondromalacia patellae. J Bone Joint Surg 43B:752–757, 1961.
19. Pai YC, Rymer WZ, Chang RW, Sharma L: Effect of age and osteoarthritis on knee proprioception. Arthritis Rheum 40:2260–2265, 1997.
20. Pap G, Meyer M, Weiler HT, Machner A, Awiszus F: Proprioception after total knee arthroplasty: A comparison with clinical outcome. Acta Orthop Scand 71:153–159, 2000.
21. Roberts D, Friden T, Stomberg A, Lindstrand A, Moritz U: Bilateral proprioceptive deficits in patients with unilateral anterior cruciate ligament reconstruction: A comparison between patients and healthy individuals. J Orthop Res 18:565–571, 2000.
22. Safran MR, Allen AA, Lephart SM, et al: Proprioception in the posterior cruciate ligament deficient knee. Knee Surg Sports Traumatol Arthrosc 7:310–317, 1999.
23. Sharma L, Pai YC, Holtkamp K, Rymer WZ: Is knee joint proprioception worse in the arthritic knee versus the unaffected knee in unilateral knee osteoarthritis. Arthritis Rheum 40:1518–1525, 1997.
24. Simmons S, Lephart S, Rubash H, Borsa P, Barrack RL: Proprioception following total knee arthroplasty with and without the posterior cruciate ligament. J Arthroplasty 11:763–768, 1996.
25. Simmons S, Lephart S, Rubash H, Pifer GW, Barrack R: Proprioception after unicondylar knee arthroplasty versus total knee arthroplasty. Clin Orthop 331:179–184, 1996.
26. Skinner HB, Barrack RL, Cook SD: Age-related decline in proprioception. Clin Orthop 184:208–211, 1984.
27. Skinner HB, Barrack RL, Cook SD, Haddad RJ: Joint position sense in total knee arthroplasty. J Orthop Res 1:276–283, 1984.
28. Skinner HB, Wyatt MP, Hodgdon JA, Conard DW, Barrack RL: Effect of fatigue on joint position sense of the knee. J Orthop Res 4:112–118, 1986.
29. Wada M, Kawahara H, Shimada S, Miyazaki T, Baba H: Joint proprioception before and after total knee arthroplasty. Clin Orthop 403:161–167, 2002.
30. Warren P, Olanlokun T, Cobb AG, Bentley G: Proprioception after knee arthroplasty: The influence of prosthetic design. Clin Orthop 297:182–187, 1993.
31. Wolfson L: Gait and balance dysfunction: A model of the interaction of age and disease. Neuroscientist 7:178–183, 2001.