Knee flexion contractures in institutionalized elderly: prevalence, severity, stability, and related variables.

Key Words: Aged, Contracture; Geriatrics, Institutionalization; Lower extremity, knee.

The development of knee flexion contractures (KFCs) in nursing home residents can limit residents' functional capabilities, can increase the nursing care needed, and often can add to the additional time and expense for physical therapy intervention. Severe KFC in the elderly can decrease potential for ambulation, and it also usually reduces the ease of independent or assisted toilet transfers needed to maintain continence. Achieving a comfortable sitting position in a standard wheelchair or a "geri-chair" may become a major problem for the resident with severe contractures. Severe contractures can also make it difficult for the nursing staff to provide adequate hygiene, skin care, and positioning for the resident.

The federal regulations (Omnibus Reconciliation Act, 1989) governing long-term care facilities specify that

.. a facility must ensure that a resident

who enters a facility without contractures

does not experience an unpredictable

reduction in range of motion

without justifiable cause, and that a

resident with contractures receives

treatment and services consistent with

professional nursing standards and

designed to increase ROM or maintain

existing ROM.[1]

Therefore, the development of contractures could represent a facility violation from the view of state survey teams.

Bergstrom and colleagues[2,3] reported a low prevalence (8%) of loss of knee extension in 79-year-old subjects, although most of these subjects were not institutionalized and showed good lower-extremity function in that they could climb stairs and walk 1,000 m. There is no similar literature describing an institutionalized geriatric population. Such information seems essential for defining the incidence of contracture development, identifying those who are at risk, and selecting those residents for whom intervention should be most intense.

The purposes of our study were (1) to describe a nursing home population in terms of factors that may be related to KFC, (2) to document the prevalence and severity of KFC in this population, and (3) to measure change in KFC and other variables over a 10-month period. The study was also designed to identify which of the variables may predict development or progression of KFC.

The variables studied were selected based on our clinical observations. In our experience, multiple problems are often observed in nursing home residents with KFC, and many of these problems seem linked to decreased mobility. Joint deformity, joint pain, hyperreflexia, and cognitive impairment could conceivably relate to this impaired mobility. It seems likely that this decreased mobility can then lead to contracture and loss of joint motion.

Method

Subjects

Prior to initiation of data collection, all residents of Villa Clement Nursing Home, Milwaukee, Wis, were asked to participate in the study. Those agreeing to participate signed an informed consent document. if a resident was unable to read, understand, or sign the document, this was done by a designated health care representative or guardian.

Villa Clement Nursing Home is a 196-bed combined intermediate care and skilled nursing facility. At the time of the study, 79% of the residents required skilled nursing care, and 21% required intermediate care. Twenty percent of the facility's residents were discharged from this nursing home to private residences over a 1-year period (data from same year as data collection for this study).[4]

Of the 196 residents of the nursing home, 112 (57%) agreed (or gave consent through a designated health care representative) to participate in the study. Of the 112 participants, 87% were female. The most frequently observed diagnostic categories among the residents who participated were organic brain syndrome (50% of the subjects), rheumatic disease (32%), cerebrovascular disease (27%), hip fracture (12%), and Parkinson's disease (4%). Fifty-one of the 112 residents were receiving physical therapy for some lower-extremity involvement at the initiation of the study. The majority (67%) of these 51 residents were receiving a maintenance level of therapy one to two times per week. The average duration of therapy was 16 months (range= 1 month to 7 years 11 months).

Although we did not compile information about those residents who did not agree to participate in the study, we are aware that the nonparticipants were residents with various levels of cognitive function. Reasons for residents refusing to participate included not wanting to be bothered, not wanting to participate without personal benefit, and lack of understanding of the research project in spite of explanation. Only a small number of residents were excluded from the study because a health care representative did not respond to calls or mailings.

Procedure

The following data were collected for each participating resident at the start of the study ([T.sub.1]) and again 10 months later ([T.sub.2]): age; gender; diagnoses; length of stay in the nursing home; and whether physical therapy was received at the time the study began and, if so, how often and over what period of time. In addition, measurements were made of ambulation level, cognitive status, passive hip and knee extension range of motion (ROM), and the angle of genu varus or genu valgus, if appropriate. Assessments were also made of the presence of pain and resistance to passive motion.

The descriptive data for each resident were retrieved from the medical chart. Specific diagnoses were grouped into major diagnostic categories (ie, diagnostic related groups). Because our study focus was on lower-extremity measurements, residents were considered as receiving physical therapy only if treatments were for a lower-extremity problem. No attempt was made to change, control, or describe the therapy programs for this study. Data on therapy were collected to describe the sample and because of the possible relationship of this intervention to change in contractures.

The nursing staff assessed the cognitive level of each resident using the Short Portable Mental Status Questionnaire (SPMSQ),[5] a 10-question test designed to measure intellectual functioning. The reliability and validity of measurements obtained with this instrument have been reported by Pfeiffer.[5] Scores on the SPMSQ fall into one of four categories: 0 to 2 errors, intact intellectual function; 3 to 4 errors, mildly impaired function; 5 to 7 errors, moderately impaired function; and 8 to 10 errors, severely impaired function.

A physical therapist, with input from the nursing staff on the clients with whom the physical therapist was not familiar, categorized each resident's ambulation level using a six-point scale devised by the authors and three colleagues involved in geriatric care (Appendix).

Over a 2-week period, within the same time frame as the above data collection, two physical therapists obtained bilateral measurements of passive hip and knee extension ROM and angle of genu varus or genu valgus for each subject. Range of motion measurements were made using a standardized procedure with a 360-degree goniometer,(*) which was covered on one side to prevent the therapist from reading a value until after the measurement was made. landmarks used for goniometric measurement (stable arm, center of rotation, and movable arm) were midlateral trunk, greater trochanter, and lateral femoral epicondyle for hip extension; greater trochanter, lateral femoral epicondyle, and lateral malleolus for knee extension; and the anterior superior iliac spine, midpatella, and the point midway between the malleoli for genu valgus.

Prior to the measurements, the resident was positioned supine in bed, with the bed completely flat. Measurements of passive hip extension were made with the opposite hip maximally flexed (as in the Thomas test) within the resident's tolerance. The resident was moved to the edge of the bed so that the limb being measured was free to extend over the side. Prior to measurements of passive knee extension, the ipsilateral hip was allowed to move into maximum extension to eliminate the effect of hamstring muscle shortness on passive knee extension ROM. The therapist passively extended the knew joint through its full ROM until maximum resistance was felt or until the resident reported pain. An assistant held the limb in the position while the therapist made the measurement. Measurements of genu varus or genu valgus were made with the knee in maximum extension. Pain during passive knee movement was recorded as present or absent, based on verbal or nonverbal expression of pain during the measurement session. The therapist also noted whether resistance to passive knee extension was present or absent throughout the ROM. Measurements made 10 months later were done by the same therapist who made the initial measurements.

Data Analysis

To describe the study population at both [T.sub.1] and [T.sub.2], statistical analysis of all variables was done using the Statistical Package for the Social Sciences-X (SPSSx) software program.[dagger] This analysis generated means, standard deviations, median values, and frequency distributions for each variable of interest. Individual subject data were also printed to allow further study and explanation of the group data. Spearman correlation coefficients were calculated to determine whether significant relationships existed between any of the variables (ROM, ambulation, cognitive status, pain, resistance, and frequency of therapy) at either measurement time. A multivariate analysis of variance (MANOVA) was done to assess the significance of the differences between [T.sub.1] and [T.sub.2]. A MANOVA and t tests were also done to compare [T.sub.1] data on those who completed the study versus those who died or dropped out of the study. Variables in the MANOVA were age, gender, ROM, ambulation, cognitive status, knee pain, resistance to passive motion, and frequency of physical therapy. Chi-square analysis was used to identify whether any of the [T.sub.1] variables might predict a change in KFC over the 10-month measurement period. Variables in the chi-square analysis were ambulation status, cognitive status, resistance, pain, and participation in a physical therapy program for 5 or more months during the 10-month study period.

Intrarater and Interrater Reliability

Intrarater and interrater reliability of goniometric measurements was calculated on randomly chosen subjects from the total study sample at [T.sub.1]. Intrarater reliability data were collected by each of the two therapists in the same morning. Ten subjects were measured and then remeasured in the same order. Intraclass correlation coefficients (CC[2,1])[6] for knee extension, knee valgus, and hip extension repeated measurements were .996, .850, and.989. Differences between repeated measurements of knee extension averaged 1 degree (range = 0 [degree]-6 [degrees]).

Interrater reliability data were collected by the two therapists in the same morning on another 10 subjects. The ICCs (2,1) were .968, .459, and .244 for knee extension, knee valgus, and hip extension. Differences between therapists on knee extension measurements averaged 3 degrees (range=0 [degree]-10 [degrees]). One therapist's measurements were not consistently higher or lower than the other's measurements. Because of the low reliability between raters on genu valgus and hip extension measurements, we did not use these ROM measures in our data analysis other than reporting means and standard deviations for the population.

Results

Initial Data Collection ([T.sub.1])

Table 1 summarizes the quantitative data collected at the initial measurement ([T.sub.1]). Ambulation scores were distributed over all six categories (Appendix). Forty percent of the residents were rated as nonambulatory, 25% were considered independent, 20% required the assistance of one person, and 15% required the assistance of two persons or the use of parallel bars in order to ambulate. The median ambulation score was 5. Based on SPMSQ scores, the subjects were classified as belonging to one of four categories: intact intellectual function (n=31), mildly impaired function (n=16), moderately impaired function (n=23), or severely impaired function (n=35). The median SPMSQ score was 5 errors. The distribution of passive knee extension measurements was skewed toward the 0-degree end of the ROM (full extension) (Tab. 2). Thirty-seven of the 112 residents (33%) had bilateral passive knee extension of 5 degrees or less (ie, lacked full extension). Only 14 subjects (12%) had full passive knee extension (ie, were able to reach 0 [degree]) bilaterally. [TABULAR DATA 1 OMITTED]

Table 2. Prevalence and Severity of
Knee Flexion Contractures ([T.sub.1])
(Percentage of Total)
Knee Flexion             Right Knee    Left Knee
Contracture ([degree])     (n=112)      (n=111)
0(a)                          21           22
1-5                           28           26
6-10                          28           28
11-20                          9            9
21-30                          5            5
31-40                          4            5
41-50                          3            2
51-100                         4            5
Total(b)                     102          102
(a)[0.sup.[degree]]=full extension.
(b)Total is greater than 100% due to rounding of
numbers.

Pain and resistance to passive motion were not present in the majority of the residents at [T.sub.1]. Pain was absent when 71% of the knees were moved (bilaterally absent in 64% of the residents). Resistance was absent in 77% of the knees (bilaterally absent in 72% of the residents). When present, both pain and resistance were more often bilateral than unilateral.

There was a significant correlation between the degree of right and left knee extension (r=.57, P<.001). The degree of passive knee extension correlated significantly with the other major variables at [T.sub.1]. A greater degree of KFC was associated with the presence of resistance to passive motion (r=.44 [right], r=.54 [left]; P<.001), greater cognitive impairment (r=.33 [right], P<.001; r=.22 [left], P<.01), impaired ambulation (r=.41 [right], r=.34 [left]; P<.001), and the presence of knee pain (r=.20 [right], P<.01; r=.43 [left], P<.001).

Impaired ambulation was also associated with impaired cognitive functioning (r=.43, P<.001) and the presence of resistance to passive motion (r=.47, P<.001). No relationship was found between pain and ambulation.

Data Collection After 10 Months ([T.sub.2])

Of the initial 112 residents in the study, 28 residents (25%) died before the 10-month remeasurement time and 2 residents were in the hospital at [T.sub.2], leaving 82 residents available for the repeated measurements. The MANOVA and t-test results comparing the residents who died or were hospitalized (n=30) and those remaining in the study after 10 months (n=82) showed no significant differences for any variable except age (the residents who died were older).

The MANOVA and t tests showed no significant differences between the [T.sub.1] and [T.sub.2] mean measurements for any variable. The mean and median ambulation scores at [T.sub.2] were the same as those obtained at [T.sub.1] The SPMSQ (cognitive function) mean and median scores increased to 6 errors at [T.sub.2], but remained in the range of moderate impairment. The absence of pain and the absence of resistance to passive motion continued to characterize the majority of the sample at [T.sub.2] as at [T.sub.1] (pain absent in 68% of the knees and resistance to passive motion absent in 74% of the knees at [T.sub.2]).

Average knee extension at [T.sub.2] was essentially unchanged compared with that measure at [T.sub.1], (right knee, 10 [degree]; left knee, 9 [degrees]). The prevalence and severity of KFC at [T.sub.2] were also similar to findings at [T.sub.1]. Of the 79 subjects available for both measurement times and for whom we had complete knee measurement data, 10 subjects (13%) had full passive knee extension (0 [degree] bilaterally at both [T.sub.1] and [T.sub.2] Twenty of the 79 subjects (25%) had bilateral passive knee extension of 5 degrees or less at both measurement times.

Stability of Measured Knee Extension Over Time

For this study, we chose to examine only differences greater than 6 degrees because our intrarater reliability data showed that differences of up to 6 degrees could occur between repeated measurements of knee extension. Table 3 shows the incidence and direction of changes in passive knee extension over the 10-month measurement period, using the 6-degree criterion for change. The majority of knees showed no change in knee extension over the 10-month study period.

Table 3. change in Passive Knee
Extension from [T.sub.1] to [T.sub2] (Frequency
Count and Percentage of Knees)
               No. of    Percentage
               Knees     of Total
No change
  >6 degrees    124          77
Gain in
  extension
  >6 degrees      22         14
Loss of
  extension
  >6 degrees      15          9
Total            161        100

The knees that gained or lost extension over the 10-month period were similar in amount of change. The mean change was 16 degrees for knees with a loss in extension (range=7 [degrees]-59 [degrees] and 12 degrees for knees with a gain in extension range=7 [degrees]-31 [degrees]). The median amount of change in knee extension was 10 degrees, regardless of the direction of change. Changes in knee extension were unilateral in most instances. Only one subject had a loss in extension in one knee and a gain in the opposite knee. The few other subjects with bilateral changes in extension showed changes in the same direction for both knees.

Chi-square analysis demonstrated that a significantly higher proportion of those knees that showed a change in KFC at [T.sub.2] also had resistance to passive motion at [T.sub.1] compared with those knees showing no change (for knees with increased KFC at [T.sub.2], [X.sup.2]=3.84, P<.05; for knees with decreased KFC at [T.sub.2], [X.sup.2. = 10.45, P<.01). The presence of knee pain at [T.sub.1] was also significantly more frequent in the knees that showed a decrease in KFC (improved extension) at [T.sub.2] ([X.sub.2] = 10.18, P<.01). Chi-square analysis did not show that ambulation status, cognitive status, or participation in a physical therapy program were significant factors in predicting either change or lack of change in knee extension measurements over the study period.

Discussion

Contracture is not a simple phenomenon. We have used the term "contracture" to describe a loss of full joint motion. Thus, in our measurement of knee extension, we took care to also extend the hip, putting slack on the hamstring muscles as they crossed the hip joint in order to minimize the influence of hamstring muscle length on knee extension. Although we used positioning to limit the influence of the antagonist muscle length on our measurements, there are other factors that may affect the measurements: pathologic muscle shortening (eg, muscle contracture, hyperreflexia: guarding due to pain); shortening or adhesions of ligaments, tendons, or joint capsules; mechanical restriction within the joint (eg, irregularity of the opposing joint surfaces or increased fluid volume in the joint); or the duration and force of the stretch before measurement.

In our study population, which we think is representative of many geriatric long-term care settings, most knees showed extension deficits of 10 degrees or less, most knees did not have pain or abnormal resistance to passive motion, and the variables measured showed essentially no change over the 10-month study period.

Compared with other data, however, our subjects' apparently small knee extension deficits may be noteworthy. Roaas and Andersson[7] reported passive knee extension ranging from 0 to 10 degrees of hyperextension (X[bar] = 1.6 [degrees]) in healthy male subjects aged 30 to 40 years. The American Academy of Orthopaedic Surgeons[8] estimates normal knee extension as 10 degrees of hyperextension, although the study population and the measurement technique used to obtain this estimate were not defined. In a study of healthy men and women aged 55 to 84 years, Smith and Walker[9] found mean passive knee extension of 0 degrees for men and hyperextension of 1 degree for women, with no significant differences among the three age subgroups in their sample. These studies of passive knee motion in healthy subjects indicate that hyperextension is normal and that loss of knee extension may be normal between the fourth and sixth decades of life, with no further significant loss up to age 84 years. Our subjects' ages fall in the upper end of or exceed the age range of Smith and Walker's sample, so some further loss in extension could be occurring in our study population related to normal aging. Our subjects, however, did not show hyperextension, and only a small percentage (13%) had bilateral full (0 [degree]) knee extension at either [T.sub.1] or [T.sub.2]. Even if "no contracture" is defined less rigidly as lacking 5 degrees or less of full passive extension, only one fourth of our subjects had "no contractures" bilaterally at both [T.sub.1] and [T.sub.2], leaving the remaining 75% with some degree of unilateral KFC.

Full knee extension is obviously not the norm in our nursing home population. Although our data showed a correlation between loss of knee extension and ambulation impairment, the functional significance of the prevalent minor decrease in knee-extension cannot be discerned from our study. Bergstrom et al[2] reported significant correlations between decreased knee ROM in 79-year-old subjects and their ability to get up from a chair, climb stairs, and enter public transportation. That study, however, did not differentiate between decreases in flexion or extension in the correlation, and it seems likely that the functional activities studied would be more affected by a decrease in flexion than by a decrease in extension. Conceivably, the prevalent small-magnitude loss of knee extension in our subjects could have ramifications for postural alignment, gait efficiency, energy expenditure, muscle forces around the joints, forces within the joint, and stress on adjacent joints.

Because we found that progressive impairment in ambulation correlated with an increasing degree of KFC and with a number of the other variables measured, we more closely examined the relationship between ambulation and KFC. A plot of ambulation scores against degree of passive knee extension (Figure) showed that KFCs of less than 20 degrees occurred in subjects across all six ambulation categories. Most of the knees with contractures of 20 degrees or greater, however, were in subjects in the nonambulatory category, and all KFCs greater than 33 degrees were in individuals who were nonambulatory ([T.sub.1]). These observations suggest that in clinical interpretation of contractures, perhaps a 30-degree KFC should be considered "severe," because it appears to be coincident with ambulation status. Therefore, perhaps a KFC approaching 20 degrees should trigger consideration of specific interventions. Obviously, nonambulation status alone does not imply severe KFC.

In further study of the nonambulatory group, chi-square analysis showed that nonambulatory subjects with KFCs greater than 20 degrees were significantly more likely to also have resistance to passive motion than were nonambulatory subjects with KFCs of less than 20 degrees ([X.sub.2[=21.60, P<.001). Although these data on contractures and ambulation status do not indicate a cause-effect relationship, it appears that the presence of resistance to motion may put a nonambulatory nursing home resident at risk for a greater KFC. Further studies are needed to confirm these relationships and to explore the effectiveness of interventions in nursing home residents who appear to be at risk.

We had expected that all or most of the variables measured would be identified as predictors or risk factors for the progression of KFC. The results of the chi-square analysis did not support this expectation, perhaps because of the small percentage of knees showing a significant change in extension over the study period. Our data do suggest that the presence of resistance to passive motion is predictive of future ROM change. Because resistance was linked with both increases and decreases in KFC, however, we are left without a clinically meaningful predictor. Most likely, the presence of resistance makes measurements less reproducible. It seems reasonable to at least conclude that when resistance to passive movement is present, therapists should rely not on one baseline measurement but on several measurements made at separate times.

Although the number of knees showing a change in extension over the study period was small, we further studied the individual data in an attempt to identify possible differences in the knees that gained extension and those that lost extension. Although the amount of change was similar for both groups of knees, the initial ([T.sub.1]) knee extension measurements were different. Of the 15 knees showing a loss in extension at [T.sub.2], 12 knees (80%) had between 0 and 11 degrees of extension at [T.sub.1] and no knees had extension deficits greater than 30 degrees at [T.sub.1]. in contrast, of the 22 knees showing gains in extension at [T.sub.2], only 5 knees (23%) had between 0 and 11 degrees of extension at [T.sub.1] and 8 knees (36%) had extension deficits greater than 30 degrees at [T.sub.1].

Chi-square analysis showed a significant difference in these initial ([T.sub.1]) knee extension data comparing the knees showing losses in extension and the knees showing gains in extension ([X.sup.2] = 9.79, P < .01). In addition, the subjects who lost extension at [T.sub.2] more frequently showed a regression in ambulation (5 subjects [42%] compared with 1 subject [6%] in the group gaining extension) and more frequently developed resistance to passive motion between [T.sub.1] and [T.sub.2] (5 subjects, compared with 1 subject in the group gaining extension). The subjects who gained extension). The subjects who gained extension at [T.sub.2] were more frequently nonambulatory at [T.sub.1] (11 subjects [61%], compared with 2 subjects [17%] in the group who lost extension). No subjects in either group improved in ambulation over the 10-month study period.

Although these numbers are too small for statistical analysis, we found it interesting that in spite of an apparent gain in knee extension in the one group, there was no corresponding improvement in ambulation. We believe that our ambulation scale was sensitive to meaningful functional change. Our data indicate that clinicians should not discount ambulatory nursing home residents with minimal KFC (0 [degree] - 11 [degrees]) as candidates for intervention, especially when resistance to motion develops and a regression in ambulation is occurring. Again, however, the effectiveness of interventions in this apparently high-risk group needs to be studied.

Our initial conceptualization of the link between decreased mobility and KFC is not unique. Other investigators[10-14] have documented the development of KFC after enforced joint immobilization in animals and humans and have described the subsequent joint and muscle changes. Evans et al[10] immobilized single knee joints in rats and reported the progressive changes that occurred in the immobilized versus the nonmobilized knee joint, mainly proliferation of connective tissue within the joint capsule, development of adhesions between the connective tissue and articular cartilage, and significant microscopic alterations and thinning of the joint cartilage. Following immobilization, the researchers stripped the musculature from the extremities of two animals and found that knee extension increased from 30 to 150 degrees. They concluded that the muscle contracture was more restrictive of knee extension than were capsular or pericapsular changes. Tabary et al[11] have demonstrated that morphological changes occur in muscle following immobilization.

Enneking and Horowitz[12] observed a sequence of changes in 10 human knees similar to that observed by Evans et al[10] in the rat knees. The humans knees were obtained for the study after above-the-knee amputations. Prolonged immobilization and marked loss of joint motion were noted in all of the knees prior to amputation, although none of the patients had diagnoses of specific knee joint disease. Enneking and Horowitz concluded that the proliferation of fibrofatty tissue around the cruciate ligaments was the major intra-articular factor limiting knee extension. Akeson et al[13] also reported an increase in intermolecular collagen cross-links in the periarticular connective tissue of immobilized rabbit knees and implicated these cross-links in contributing to the contracture process.

Although none of the subjects in our study underwent enforced immobilization of the knee joint, we did have a population in which 40% were nonambulatory and another 35% were dependent on the assistance of another person(s) to ambulate. In both of these groups, therefore, we know that the daily mobility and weight bearing of the knee joints were greatly decreased, especially in the nonambulators.

Conclusions

Although our knee extension data were skewed toward 0 degrees (full extension), comparison with other data suggests that our sample shows a high prevalence of unilateral knee extension deficits. Our study does not show a clear impact of decreased extension on function. In our sample, a 30-degree loss of knee extension appears to be coincident with loss of ambulation ability, although nonambulation does not imply loss of knee extension. The presence of resistance to passive motion throughout the range may make measurement less reproducible and indicates a need for repeated baseline ROM measurements.

Because only a small group of knees showed a change in extension over the 10-month period, no strong predictors emerged from our data regarding risk for progression of KFC. Based on our data, however, several groups of nursing home residents may merit physical therapy assessment and consideration for intervention: nonambulatory residents with resistance to passive motion, residents with KFC approaching 20 degrees, and ambulatory residents with minimal KFC who develop resistance to passive motion and begin to regress in ambulation. Further study is needed to clearly identify the impact of KFC on multiple measures of function and to test the effectiveness of interventions.

Acknowledgments

We thank Rose Fritz, PT, Denise Bibis, PT, the nursing and physical therapy staff of Villa Clement Nursing Home for their assistance in the data collection, and Sue Loeffler for manuscript preparation.

References

[1] Medicare and Medicaid: requirements for long-term care facilities (42 CFR 483.25). Federal Register. February 2, 1989;54:5334. [2] Bergstrom G, Aniansson A, Bjelle A, et al. Functional consequences of joint impairment at age 79. Scand J Rehabil Med. 1985; 17:183-190. [3] Bergstrom G, Bjelle A, Sorensen LB, et al. Prevalence of symptoms and signs of joint impairment at age 79. Scand J Rehabil Med. 1985; 17:173-182. [4] Wisconsin Nursing Home Directory Fact Book 1986. Department of Health and Social Services, Division of Health, Center for Health Statistics, Resource Data Section; 1987:82. [5] Pfeiffer E. A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. J Am Geriatr Soc. 1975;3:433-441. [6] Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420-428. [7] Roaas A, Andersson GBJ. Normal range of motion of the hip, knee and ankle joints in male subjects, 30-40 years of age. Acta Orthop Scand. 1982;53:205-208. [8] American Academy of Orthopaedic Surgeons. Joint Motion: Method of Measuring and Recording. New York, NY: Churchill Livingstone Inc; 1983. [9] Smith JR, Walker JM. Knee and elbow range of motion in healthy older individuals. Physical and Occupational Therapy in Geriatrics 1983;2(4):31-38. [10] Evans EB, Eggers GW, Butler JK, Blumel J. Experimental immobilization and remobilization of rat knee joints. J Bone Joint Surg [Am]. 1960;42:737-758. [11] Tabary JC, Tabary C, Tardieu C, et al. Physiological and structural changes in the cat's soleus muscle due to immobilization at different lengths by plaster casts. J Physiol (Lond). 1972; 224:231-244. [12] Enneking WF, Horowitz M. The intraarticular effects of immobilization on the human knee. J Bone Joint Surg [Am]. 1972;54: 973-985. [13] Akeson WH, Amiel D, Mechanic GL, et al. Collagen cross-linking alterations in joint contractures: changes in the reducible cross-links in periarticular connective tissue collagen after nine weeks of immobilization. Connect Tissue Res. 1977;5:15-19. [14] Akeson WH, Amiel D, Woo SL-Y. Immobility effects on synovial joints: the pathomechanics of joint contracture. Biorheology. 1980;17: 95-110.

Commentary

Mollinger and Steffen are to be complimented on their study of knee flexion contractures in an older nursing home population. Their results are timely, important, significant, and immediately clinically applicable.

Measures used in this study are practiced routinely by all physical therapists. One of the strengths of this investigation is that it was conducted using easily reproducible measures that almost any physical therapist can use to further our understanding of contractures. It is hoped that this study will cause physical therapists to realize that good research can be accomplished without fancy equipment. It is further hoped that this study will provide an impetus for additional new investigation.

In a typical nursing home setting, physical therapy is responsible for the functional well-being of all the residents--a large responsibility. There is never enough time in the day to provide significant care to everyone. Mollinger and Steffen's results seem to indicate that, in the majority of cases, normal movement is sufficient to maintain knee range of motion (ROM). Perhaps physical therapists do not need to worry at all times about all nursing home residents. Results do suggest that patients experiencing an episode of joint pain, those on the threshold of change in ambulation status, and those with a noticeable increase in resistance to passive motion should be given a higher priority for physical therapy intervention.

Research invariably raises more questions than it answers, and this study is no exception. The authors' responses to the following comments and questions would be welcomed. 1. Do you feel nursing care was a

factor affecting the results? Did the

nursing staff routinely reposition

residents, stand them frequently at

bedside, or give them passive

range of motion? Were residents

encouraged to ambulate? For example,

were residents walked to

meals or given the opportunity to

ambulate with assistance throughout

the day? 2. In this study, measurement of knee

valgus and varus ROM was not

reliable. Do you feel it is important

to include varus or valgus as a

variable in future studies? Would it

be reasonable to simply note the

presence or absence of varus or

valgus when a resident is standing? 3. Do you have any sense of whether

a decrease in knee ROM preceded

the change in ambulation status or

vice versa? 4. Do you have a sense of whether

ROM or resistance to passive motion

changed precipitously in those

residents who died? 5. Those residents who had gained

ROM at T2 were more frequently

nonambulatory, which is difficult to

understand. Did these residents

perhaps spend more time in bed

with their knees straight? 6. Do you feel there are other measures

of function (eg, transfers)

that should be addressed in future

studies? 7. Is there a particular reason why

10 months was chosen as the [T.sub.2]

measurement?

In summary, this study has provided physical therapy with a better understanding of factors placing older patients at risk for serious change and for selecting residents for whom physical therapy intervention should be most intense. I am looking forward to future studies.

Marybeth Brown, PhD, PT 452 Parkland Ave Glendale, MO 63122-4712

Author Response

We appreciate Dr Brown's positive comments on the clinical applicability of our study. We, the authors, are clinicians first and acknowledge that our study has modified our own practice in relation to nursing home residents with knee flexion contracture (KFC).

We have addressed the commentator's questions in order:

1. Nursing care is a factor that we did

not measure or control but that

certainly could have affected the

low incidence of increased contractures

in our subjects, We do not

think this was a strong contributing

factor for reasons that follow. Nevertheless,

we reiterate our belief

that our study site is representative

of high-quality homes with good

nursing care. That is, daily nursing

care includes repositioning to

prevent the formation of decubiti

when residents are in bed (ie,

turning, but usually not consistently

placing rolls or pillows that might

be thought to prevent contractures),

assisting with ambulation

within the resident's room (ie, to

the bathroom), and encouraging

ambulation to the dining room for

residents living on the same level

and near the common dining

room. Daily nursing care most

likely does not include passive

range of motion exercise as understood

and taught by physical therapists;

assistance with standing at the

bedside for the sake of weight

bearing; or routine opportunities

to ambulate with assistance outside

of a resident's room throughout

the day, especially for residents

who are not within a reasonable

walking distance of the dining

room.

Future investigations of contracture development and management could add important clinical information by recording frequency of weight-bearing activities in this population and its relationship to contractures. 2. We originally included measurement

of genu varus/valgus as an

indicator of joint deformity. As we

learned, this is not a very reliable

measure in the presence of significant

contractures. As suggested,

reporting of presence/absence

along with indication of laxity may

be more clinically valuable information. 3. Our data do not provide a clear

indication of whether a decrease in

range of motion (A) preceded a

decline in ambulation (B) (if A,

then B) or vice versa (if B, then A).

The fact that, at both the initial

measurement ([T.sub.1]) and the

10-month remeasurement ([T.sub.2]),

nonambulators were among those

with KFC of less than 20 degrees

confirms that decline in ambulation

ability is not always associated

with significant KFC nor does nonambulation

necessarily lead to

severe KFC. Thus, the proposed "if

B, then A" is definitely not always

true, The fact that all residents with

a KFC of greater than 33 degrees

were nonambulatory suggests that

the notion of "if A, then B" could

be true. But, again, our data do not

prove a cause-effect relationship.

More likely, multiple factors decide

the flow of regression. For example,

we have observed and worked

with residents with long-standing

histories of arthritic knee pain who

appear to have an almost simultaneous

decline in ambulation and

increase in KFC. In these cases, the

worsening pain (often in combination

with some cognitive impairment)

seems to trigger the downward

spiral of decline in both

ambulation and knee extension. 4. We know nothing more specific

about the residents who died than

that they were not different from

the others in any variable at [T.sub.1],

except that they were older. We

cannot comment on any change in

resistance in motion in this group

with only this one measure. Our

clinical experience suggests that

there is no sharp change in range

of motion or resistance prior to

death. 5. It is unlikely that the gain in range

in motion at [T.sub.2] for some residents

can be explained by their spending

more time in bed with their knees

straight. As mentioned before, our

experience suggests that in reality,

residents are not positioned in bed

for contracture prevention but

rather for decubitus prevention.

Focusing on the increased frequency

of nonambulators in the

group who appeared to gain range

of motion at [T.sub.2] is probably misleading.

Residents in this group

were also more likely to have

more severe contractures at [T.sub.1] and

to have knee pain at [T.sub.1]. Most likely

their nonambulatory status was

linked to their already severe KFC.

One is still left to wonder why this

initially more "impaired" group

should show an improvement in

knee extension at [T.sub.2]. In looking at

diagnosis as a possible explanation,

we found it interesting that all four

subjects with Parkinson's disease in

the study were in the groups with

a change in KFC. Parkinson's disease

and hip fractures were the

only two diagnoses that showed an

uneven split between the two

groups with a change in KFC (ie,

three subjects with Parkinson's

disease showed a gain, and one

showed a loss in range of motion;

none of the subjects with hip fracture

showed a gain, and three

showed a loss in range of motion).

Subjects in both of these diagnostic

groups showed some of the largest

changes in the study (10 [degrees]-59 [degrees]).

Might there be fluctuation in the

knee extension measurement of

these patients from day to day

based on factors other than true

contracture? Although this finding

was interesting, these subjects

accounted for only a small number

of the knees showing change. 6. Future studies would clearly be

strengthened by reporting multiple

measures of function: ambulation,

transfers, bed mobility, and ability

to be positioned in a chair or bed.

Other useful measures would be

restraint use and the frequency of

weight-bearing activities. 7. In choosing a time frame for the

second measurement, we considered

both a time sufficient for

change to occur and a reasonable

duration to commit to a clinical

research project. We felt that 6

months was a minimum time for

expecting to see measurable

change; 10 months happened to be

what fit into the work schedules of

our measuring therapists. (*) Fred Sammons Inc, PO Box 32, Brookfield, IL 60513-0032. (dagger) SPSS Inc, 444 N Michigan Ave, Chicago, IL 60611.