Unloader Knee Braces for Osteoarthritis: Do Patients Actually Wear Them?

Abstract

Background

Unloader braces are a nonsurgical approach for predominantly unicompartmental knee arthritis. Although noninvasive, braces are expensive and it is unclear whether clinical factors, if any, will predict regular brace use.

Questions/purposes

We asked: (1) Do patients continue to use the unloader brace more than 1 year after it is prescribed? (2) Do any clinical or radiographic factors predict continued use of the unloader brace after the first year? (3) What are the most common subjective reasons that patients give for discontinuing the brace?

Methods

We administered 110 surveys to all patients who were fitted for unloader knee braces for predominantly unicompartmental osteoarthritis 12 to 40 months before administration of the survey. Standardized indications and fitting protocols were used. The following parameters were tested for association with ongoing brace use: alignment, arthritis severity, compartment involved, BMI, weight, age, gender, pain and function, number of refittings, and problems with the brace. The survey response rate was 81% (89 of 110).

Results

Of the 89 responders, 28% reported regular brace use (twice per week, an hour at a time, or more); at 2 years, 25% used the brace regularly. No clinical or radiographic factors considered were associated with ongoing brace use. Patients reported lack of symptomatic relief, brace discomfort, poor fit, and skin irritation as reasons for discontinuing the brace.

Conclusions

Surgeons and patients need to balance the benefits and absence of complications of bracing against cost and the low likelihood of ongoing use 1 year or more after the prescription of the brace.

Level of Evidence

Level III, prognostic study. See Instructions for Authors for a complete description of levels of evidence.

Introduction

Knee arthritis is common, underreported, and increasing in prevalence [1, 3, 4, 15]. In 2009, surgeons performed 686,000 knee arthroplasties [13]; projections going forward predict 1.52 million procedures in 2020 and 3.48 million procedures in 2030 [13]. The revision burden, likewise, continues to mount; a 600% increase in revision procedures is expected by 2030 [13]. With numbers like these, any promising nonsurgical approach deserves careful evaluation.

NSAIDs, injections, and narcotic analgesics are in common use, but each of these has disadvantages. The use of NSAIDs frequently is limited by gastrointestinal and renal side effects, especially in the elderly [19]. A recent meta-analysis investigating the benefit of injections concluded intraarticular steroid injections could provide pain relief for knee osteoarthritis up to 1 year and hyaluronate was beneficial specifically in patients who are poor surgical candidates and have mild radiographic disease [7].

Arthritis unloader braces provide a low-risk intervention for selected patients. One randomized clinical trial compared medical management, neoprene sleeves, and valgus unloader knee braces, and found improved disease-specific quality of life and function scores in patients using neoprene sleeves and in those using unloader braces, but only trends toward improvement in aggregate WOMAC^TM scores compared with patients who received medical management alone [12]. In addition, patients wearing unloader knee braces had reduced pain compared with patients wearing neoprene knee sleeves in 6-minute walking and 30-second stair-climbing tests [12]. In a selected literature review on knee bracing for unicompartmental osteoarthritis, Pollo and Jackson [16] cited studies showing improvement in pain scores with brace wear, improved stride symmetry, and a decreased varus moment in valgus-braced knees. No clinical studies regarding brace compliance were available for review at the time of that report, but the authors cited compliance greater than 75% in their own clinic population. The authors concluded, “Evidence supports the clinical efficacy of bracing for managing osteoarthritis of the knee.” In contrast, the 2009 the American Academy of Orthopaedic Surgeons Clinical Practice Guideline for knee arthritis concluded that the literature provided insufficient evidence to recommend for or against the use of varus- or valgus-producing unloader knee braces [18].

Several studies on brace compliance differ in terms of whether patients continue to use the unloader brace with time. One study in a Veterans Affairs population evaluated brace use in 48 men and one woman with an average age of 57 years and found 76% of patients used the brace at least once weekly at 1 year and 69% at 2 years [10]. Another study reported improvements in pain relief in patients wearing the brace longer than 8 weeks, but despite this relief, compliance with brace use at a mean of 2.7 years was only 42% [2]. That same group of patients was recontacted in a followup study; none continued to use the brace at 11 years [21]. A Cochrane review from 2005 concluded that braces have some (albeit limited) efficacy for treatment of medial compartment osteoarthritis and little long-term use [5]; an additional study in 2006 found 42% of subjects discontinued brace use within the 12-month trial period, 88% of whom discontinued the brace within the first 6 months [6]. Given the disparate findings pertaining to ongoing brace use—from 0% to 76%— we sought to look specifically at patient use (or lack of use) of unloader braces. Patient compliance is important because of the cost of these braces; although unloader braces do not carry much risk (there is a single case reported associating pulmonary embolism and deep venous thrombosis with brace use [10]), these braces are expensive. The patient charge for one kind of brace is nearly USD 1800 at our institution. By comparison, the professional fee for a knee arthroplasty for a Medicare patient is approximately USD 1450 [9].

We therefore asked three study questions: (1) Do patients continue to use the unloader brace more than 1 year after it is prescribed? (2) Do any clinical or radiographic factors predict continued use of the unloader brace after the first year? (3) What are the most common subjective reasons that patients give for discontinuing the brace?

Patients and Methods

Surveys were administered to all patients who underwent initial fitting for an unloader knee brace from October 2007 to June 2010. All patients initially seen for knee OA during the study period regardless of severity were offered an unloader brace (Össur Americas, Foothill Ranch, CA, USA) as part of the nonoperative approach. The indications for brace use were predominantly medial or lateral radiographic osteoarthritis, or pain clinically correlating with the most involved compartment on weightbearing radiographs. The contraindications for brace prescriptions were superficial wounds, arterial insufficiency, or severe varicosities that could result in skin at risk with regular brace wear. Weight, BMI, radiographic severity, and limb alignment (varus or valgus) were not used as contraindications to offering patients the brace as an option during this period. Patients were free to decline the brace prescription. We sent postal surveys to all 110 patients fitted for an unloader brace during the period in question; patients could actively opt out of any specific question or the study in entirety by mail, over the phone, or simply by nonresponse. Brace fitting occurred at a mean 24 months (range, 12–40 months) before the survey was mailed. The survey response rate was 81% (89 of 110). The mean age for survey responders was —63 years (SD, 9.4 years), and the mean BMI was 28 (SD, 5.6). The population was 52% male and 48 % female; 59% had predominantly varus alignment and 68% had predominantly medial compartment involvement (Table 1).

Table 1.

Demographics of survey responders

Variable	Value
Age (years)*	63 ± 9.4 (43–83)
BMI*	28 ± 5.6 (20–50)
Weight (kg) *	86 ± 20 (50–145)
Sex
Male	52%
Female	48%
Radiographic alignment
Varus	59%
Valgus	26%
Neutral	15%
Most involved compartment
Medial	68%
Lateral	32%
Arthritis severity (1–3)**
Medial compartment	2.39 ± 0.86
Lateral compartment	1.66 ± 0.88
Patellofemoral compartment	1.26 ± 0.49

* Values are expressed as mean ± SD, with range in parentheses; ** modified Kellgren-Lawrence scale as follows: Grade 1, minimal to no osteophytes or joint space narrowing; Grade 2, osteophytes and/or moderate joint space narrowing (Kellgren-Lawrence Grade 1 or 2); and Grade 3, severe joint space narrowing (Kellgren-Lawrence Grade 3 or 4).

The same surgeon (SSL) wrote all the brace prescriptions. All braces were fitted by the same experienced orthotist, who fits approximately 160 braces per year and has worked clinically on a referral basis with the senior investigator (SSL) since 2007. The orthotist is not associated with the referring institution or the referring surgeon in any other way. All patients were fitted using the same family of braces and the same guiding principle for proper fit: neutral alignment or the smallest amount of unloading that generated a level of symptomatic relief that the patient found satisfactory, whichever correction was smaller. Off-the-shelf braces were used when possible; custom-fit braces were used when necessary, usually based on the size or alignment of the limb. All patients were instructed to return to the orthotist as needed for refittings without charge. All patients were offered a trial of the brace and were encouraged to return it at no cost within 30 days for any reason. Of the 110 patients who received a survey, 88 patients (80%) had been fitted for a custom-fitted unloader brace and 22 (20%) for an off-the-shelf brace.

We encouraged patients to use various nonsurgical approaches; the menu of nonoperative cotreatments we used included infrequent use of NSAIDs and nonnarcotic analgesics, a recommendation against narcotic analgesics, occasional use of intraarticular corticosteroid injections and/or viscosupplementation injections, assistive devices for ambulation when necessary, weight control, and reasonable activity modifications. In our study, we sought neutral alignment or the smallest coronal-plane correction that provided symptomatic relief. This approach is supported by the work of Ramsey et al. [17] who evaluated the mechanics of brace efficacy in 16 patients with medial compartment osteoarthritis; that group used gait analysis and compared bracing with neutral alignment to an overcorrection of 4° valgus and found the smaller correction to result in better pain scores, knee function, and gait.

The survey asked about current and past brace use, knee pain, function, satisfaction with the brace, and problems related to use of the brace, such as skin irritation or knee swelling. The mailed cover letter and survey were intentionally designed to minimize bias in responses. (Appendix 1). We defined regular brace use as at least an hour a day, at least 2 days a week. Two weeks after the survey was mailed, patients who did not respond to the survey were contacted by telephone. All patients were given the option of declining to participate and could do so by telephone, by email, or by returning a postage-paid form indicating their desire to be excluded from the study.

We determined height, weight, age, BMI, and sex by chart review. We determined the severity and pattern of knee arthritis by radiographic review; arthritis severity in the most-involved compartment was one variable considered, but we also investigated the potential impact of radiographic changes in the less-involved (or uninvolved) compartments. Radiographs were independently analyzed by two reviewers (ES, DLS); discrepancies were adjudicated when necessary by a third reviewer (SSL). Arthritis severity was determined using standing AP and Rosenberg views and a patellar sunrise view. Arthritis was graded in the medial, lateral, and patellofemoral compartments using a modification of the Kellgren-Lawrence grading system [11] as follows: Grade 1, minimal to no osteophytes or joint space narrowing; Grade 2, osteophytes and/or moderate joint space narrowing (Kellgren-Lawrence Grade 1 or 2); and Grade 3, severe joint space narrowing (Kellgren-Lawrence Grade 3 or 4). Limb alignment was characterized as varus, valgus, or neutral on standing films.

Descriptive statistics of available responses and data were used for primary data explorations and comparisons. Logistic regression was used to determine whether respondent attributes (BMI, weight, age, sex, radiographic severity, limb alignment) were associated with the likelihood of using a brace for 3, 6, and 12 months. A two-tailed Fisher’s exact test was used to determine whether a difference in brace use was associated with an increased maximum walking distance. We performed statistical analysis using PASW^® Statistics 18 (IBM Corp, Armonk, NY, USA).

Results

Only 25 (28%) of the 89 patients continued to use the brace regularly more than 1 year after brace fitting (Table 2); mean followup for these patients was 24 months (range, 13–40 months). Of the 40 patients receiving a brace more than 2 years before the survey, 10 (25%) reported continued regular brace use. Of the 14 patients receiving a brace more than 3 years before the survey, three (21%) reported continued regular brace use. Nine of 89 patients (10%) who did not meet our definition of regular brace users reported more occasional brace use (once per week, an hour at a time). Most patients did not use their braces much, if at all, beyond the 1-month trial period: of the patients who discontinued use of the brace at less than a year, 26% reported never having used the brace regularly, and 39% did so for less than 3 months. In aggregate, then, 65% of the patients fitted for braces did not use them even for 3 months.

Table 2.

Brace use through time (months since brace fitting)

Months since fitting	Number of surveys completed	Number of braces in use
> 12	89	25 (28%)
> 24	40	10 (25%)
> 36	14	3 (21%)

None of the potential predictor variables we analyzed (BMI, weight, age, sex, radiographic severity, limb alignment) were associated with brace use or discontinuation. Several subjective parameters from the survey were associated with brace use. Patients who described at least a considerable improvement in walking range (defined as more than twice the walking range than without the brace) were more likely (p < 0.001) to continue to use the brace beyond 1 year. By contrast, patients who enjoyed smaller subjective improvements in function (less than doubling of walking range) were more likely to discontinue brace use before 1 year (Table 3).

Table 3.

Association of use longer than 1 year with improved walking range

Walking range improved greater than twice the range without the brace?	Number of patients		p value
	Brace used at least 1 year?
	No	Yes
No	23 (68%)	7 (24%)	< 0.001
Yes	2 (5.8%)	22 (76%)

Nine patients reported not having used their brace enough to comment; remaining surveys (26) submitted with this question omitted.

Patients cited numerous subjective factors that led to brace discontinuation. Those who did not wear the brace for 1 year were more likely to describe difficulties with the brace than patients who continued to use the brace. Problems patients reported in the survey included skin irritation/swelling, poor fit, lack of symptomatic relief, difficulty donning/doffing brace, difficulty wearing with clothing, and heaviness/bulkiness of brace (Table 4). All patients were invited to return to have the brace readjusted as many times as needed at no charge; interestingly, of the patients who required two or more fittings, only 23% (three of 13) used the brace for a full year. By contrast, of the patients who needed the brace readjusted only once or not at all, 43% (26 of 60) used the brace for a full year. This suggests some patients may be easier to fit than others; however, we were unable to identify any factors associated with a patient who would find the brace helpful.

Table 4.

Self-reported aspects of brace that prevented greater use

Brace aspect	Number of patients
	Brace used at least 1 year?
	No (n = 42)	Yes (n = 20)
The brace causes skin irritation or swelling	17 (40%)	4 (20%)
The brace does not fit well enough or was too uncomfortable	25 (60%)	2 (10%)
The brace does not help my symptoms enough to make it worth wearing	21 (50%)	2 (10%)
The brace is hard to put on/take off	7 (17%)	0
The brace is too hard to wear with the clothes that I wanted to wear for the activities I wanted to do	13 (31%)	6 (30%)
The brace is too heavy or too bulky	14 (33%)	4 (20%)

Remaining surveys (27) submitted with this question omitted.

Discussion

Unloader braces are a possible nonsurgical approach for predominantly unicompartmental knee arthritis. Although noninvasive, braces are expensive, and little is known about which clinical factors, if any, are associated with regular, ongoing brace use. We therefore asked three study questions: (1) Do patients continue to use the unloader brace more than 1 year after it is prescribed? (2) Are there clinical or radiographic factors that predict continued use of the unloader brace after the first year? (3) What are the most common subjective reasons that patients give for discontinuing the brace?

Our study has some limitations. First, our study is subject to recall bias in terms of the subjective elements of the survey, as it required patients to recall the reasons for discontinued use of a brace that was prescribed anywhere from 6 to 38 months before. However, the survey should provide reasonably accurate appraisal of how many patients continued to use the brace and, for those who did not continue use, approximately when they discontinued its use. Second, we were unable to measure femorotibial angle on available weightbearing radiographs; hip-to-ankle radiographs were not available for all patients, as we are the referral center for five states, and patients come with outside radiographs taken with variable techniques. However, all patients had weightbearing views, and we were able to classify the patients according to broad classifications of alignment (varus, neutral, and valgus), and the Kellgren-Lawrence scale provided an established and generally accepted approach for grading osteoarthritis severity [11]. Third, we had no control or comparison group. This limitation would be more important in a study on brace efficacy; it is less critical in an “open label” study on brace compliance. We believe that there is a placebo effect at work here; patients knew they were receiving an intervention, and they had agreed to it. The 1-year compliance rate of 28% effect was considered; one surgical trial resulted in a placebo-related improvement of 37% in patients with angina pectoris [8], and an even larger subset of patients reported a response to sham arthroscopic surgery [14].

Our finding of only approximately one in four patients continuing to use the brace regularly for more than a year may represent a best-case scenario, for the brace, in that we used only one experienced orthotist who communicated regularly with the prescribing surgeon, gave all patients a no-charge 1-month trial with the brace, invited the patients to have free refittings as needed, and permitted cotreatments (including joint injections and nonnarcotic analgesics) as desired by the patients. We surmise patients who declined to participate in the survey were more likely to be unhappy with the brace treatment, as there are data to suggest health states of nonresponders are lower than those of responders in survey studies [20], again emphasizing, if anything, our results represented a best-case estimate of success with the unloader brace. Other studies have varied in terms of reported compliance rates with time. One followup survey of patients previously prescribed braces reported that 41% of 30 patients were using the brace at 2.7 years, where use was defined as an average of 5 hours per day for work or weightbearing activity [2]; when those patients (n = 29) were resurveyed at an average of 11.2 years followup, none were using the brace [21]. Another study, in a population of military veterans, found 76% of patients were still using the brace at least once a week after a year [10]. It is difficult to account for the variation among the studies. Our study used a stricter definition of brace use (twice a week, compared with once a week) than the others, and perhaps access-to-care issues associated with the managed-care population accounted for the high estimate in the Veterans Affairs study; our patients generally were commercially insured or insured through Medicare. Perhaps patients in our study believed they had more treatment choices and shorter waiting times for care than patients in a Veterans Affairs system, making them more likely to self-discontinue brace use; however, this explanation is speculative. In any case, the study in veterans [10] offered the highest estimate of ongoing use that we found in the literature.

We were unable to identify any patient or radiographic factors, such as sex, age, BMI, limb alignment, or arthritis severity that predicted use or discontinued use of the brace. Our patient population is in many ways similar to other populations studied in the brace literature in terms of age, BMI, and radiographic arthritis severity (Table 5). Our sample size of patients prescribed braces is larger than samples in the other published studies on the subject. The study by Giori [10], like ours, concluded that continued brace use was not associated with weight, BMI, or the primary compartment affected by osteoarthritis, although Giori did find that patients younger than 50 years had better brace compliance than patients older than 65 years. Brouwer et al. [6] claimed a nonsignificant trend toward better patient function and pain relief with unloader braces in younger patients. It is possible, therefore, that differences in the patient population studied—in this instance, age— may affect compliance with brace use, but the data are far from conclusive on this point.

Table 5.

Comparison of published populations undergoing unloader bracing

Study	Sample size (number of patients)	Age (years)	BMI	Arthritis severity	Brace compliance	Followup (years)
Giori [10]	46	57	32	2.96	76%, 68%, & 61%	1, 2, & 3
Kirkley et al. [12]	41	59	≤ 35*	3.39	NA	0.5
Brouwer et al. [6]	60	59	28	NA	58%	1
Barnes et al. [2]	30	57	29	2.84	41%	2.7
Wilson et al. [21]	29	66	29	3.06	0%	11.2
Current study	89	63	28	2.88**	28%, 24%, & 21%	1, 2, & 3

* Average BMI not listed; however, BMI > 35 was an exclusion criteria for this study; ** arthritis severity of current data was obtained using a modified Kellgren-Lawrence grade as described and normalized here to a 4-point scale for comparison.

Not surprisingly, patients in our study who reported substantial improvement in comfortable walking range (doubling of that range or more) and fewer difficulties with the brace, such as skin irritation or difficulties using the brace with clothing, were more likely to continue using the brace. Brouwer et al. also noted cited lack of effect (15/25), skin irritation (2/25), and poor fit (2/25) as reasons for discontinuation of brace use before 12 months [6].

Our results may serve as hypothesis-generating pilot data for a prospective trial to evaluate clinical- and/or cost-efficacy of unloader braces. Insofar as these braces are expensive, we believe that such a study is warranted; charges to patients’ insurance for the braces used in this study were USD 849 for an off-the-shelf single-hinge unloader knee brace, and USD 1780 for a custom single-hinge unloader knee brace. We now counsel our patients who are considering use of an unloader brace that the likelihood they will still be using the brace is approximately one in four. We continue to offer the unloader brace as part of a comprehensive approach to nonsurgical treatment of patients whose arthritis pattern is predominantly unicompartmental; however, our findings make us less sanguine that the brace will serve as a durable intervention, and we are candid with our patients about the relatively low likelihood that a brace, once fitted, will remain in service beyond the first year of use.

Appendix 1. Study Information Letter Survey