HHS Public Access

Author manuscript
Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.
Author Manuscript

Published in final edited form as:

Arthritis Rheumatol. 2019 June ; 71(6): 908–915. doi:10.1002/art.40808.

The Efficacy of a Lateral Wedge Insole for Painful Medial Knee

Osteoarthritis after Prescreening: A Randomized Clinical Trial
David T. Felson, MD, MPH1,2,3, Matthew Parkes, BSc (Hons)2,3, Suzanne Carter, BSc2,3,
Anmin Liu, PhD4, Michael J. Callaghan, PhD2,3,5,6, Richard Hodgson, MB BS PhD2,6,
Michael Bowes, PhD7, and Richard K. Jones, PhD4
1Clinical
Epidemiology Research & Training Unit, Boston University School of Medicine, Boston,
Author Manuscript

MA, USA.
2Arthritis
Research UK Centre for Epidemiology, Centre for Musculoskeletal Research, Faculty of
Biology, Medicine and Health, The University of Manchester, Manchester, UK.
3NIHR Manchester Musculoskeletal Biomedical Research Centre, Manchester University
Hospitals NHS Foundation Trust, Manchester Academic Health Science Centre, Manchester, UK.
4School of Health Sciences, University of Salford, Salford, UK.
5Facultyof Health, Psychology, and Social Care, Department of Health Professions, Manchester
Metropolitan University, Manchester, UK.
6Manchester University Hospitals NHS Foundation Trust, Manchester, UK
7Imorphics, Ltd, Manchester, UK
Author Manuscript

Abstract
Objective: Lateral wedge shoe insoles decrease medial knee loading, but trials have shown no
effect on pain in medial knee osteoarthritis (OA). However, insoles’ loading effects are
inconsistent, and they can increase patellofemoral loading. We hypothesized that insoles would
reduce pain in preselected patients.

Methods: In persons with painful medial knee OA, we excluded those with patellofemoral OA
and those with pain <4/10. We further excluded participants who, in a gait laboratory using lateral
wedges, did not show at least a 2% reduction in knee adduction moment (KAM) compared with
their shoes and a neutral insole. We then randomized subjects to lateral wedge vs. neutral insole
for 8 week periods separated by an 8 week washout. Primary outcome was knee pain over the past
Author Manuscript

week (0–10) and secondary outcomes nominated activity pain and KOOS pain. We carried out
mixed model analyses adjusted for baseline pain.

Results: Of 83 participants, 21 (25%) were excluded because of insufficient reduction in KAM.

Of 62 included, mean age was 64.2 years (SD 9.1); 37.1% were women. Lateral wedge insoles
produced a greater reduction in knee pain than neutral insoles (difference 0.7 on 0–10 scale;
95%CI 0.1, 1.2; p = 0.02). Secondary outcomes showed mixed findings.

Address correspondence to Dr. Felson at 650 Albany Street, X-200, Boston University School of Medicine, Boston, MA 02118. T:
617-638-5180. F: 617-638-5239. dfelson@bu.edu.
 Felson et al. Page 2

Conclusions: In persons prescreened to eliminate those with patellofemoral OA and

Author Manuscript

biomechanical non-responders, lateral wedge insoles reduced knee pain, but the effect of treatment
was small and is likely of clinical significance in only a minority of patients. Targeting patients
may identify those who respond to this treatment.

Roughly 12% of persons aged 60 and over have painful knee osteoarthritis (OA) (1). Rates
of knee replacement have been rising in large part because of the failure of medical and
rehabilitative treatments. New treatments for knee OA are badly needed.

Lateral wedge insoles placed inside shoes shift loading across the knee laterally during
walking. They reduce load across the medial knee where most affected persons have either
isolated disease or disease combined with involvement of the patellofemoral joint. Lateral
wedge insoles reduce the external knee adduction moment (KAM), a measure of the load
across the medial vs. lateral compartment, by 5–6% (2, 3). Unfortunately, in trials, lateral
Author Manuscript

wedges have not reduced knee pain compared with neutral insoles. In a meta-analysis by
Parkes et al. (4), all eight randomized controlled trials comparing lateral wedge to neutral
insoles were null, and the effect size on pain reduction was 0.03 (95% confidence interval
−0.18, 0.22). A subsequently published trial also was negative (5).

In 25% of patients, the wedge does not reduce medial load (6, 7). Furthermore, OA in the
patellofemoral joint may get worse as load is shifted laterally. We therefore hypothesized
that if we selected persons with painful medial compartment knee OA who showed a
biomechanical response to wedge insoles and did not have painful patellofemoral OA, they
would experience a reduction in knee pain compared with neutral insoles.

Because use of a shoe insole is a simple, low-cost intervention, its efficacy in reducing pain
could translate into large public health benefit and possibly widespread use. Further, medial
Author Manuscript

knee OA is highly prevalent not only in Western countries but in developing ones where
knee replacements are not widely available. We conducted a crossover trial testing a 5°
lateral wedge insole, the same insole we and others tested previously (8).

METHODS
This was a randomized trial testing lateral wedge insoles vs. neutral insoles in persons with
painful medial knee osteoarthritis. The clinical trial registration number was
ISRCTN55059760.

Recruitment and Eligibility

Recruitment—Subjects were recruited from general practices and by way of
Author Manuscript

advertisements in Manchester, England from 1/2016 through 6/2017

Inclusion Criteria—We instituted the following eligibility criteria: age 40–85; severity of
overall knee pain in the last week of >4 on a 0–10 numerical grading scale (overall knee pain
is the primary outcome of the trial), Kellgren & Lawrence grade 2–4 in the painful knee (as
scored by a musculoskeletal radiologist) on a PA or AP radiograph obtained within the last 2
years that showed definite medial but no definite lateral narrowing. Patellofemoral

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 3

osteoarthritis must be less severe than medial OA and could not be Kellgren & Lawrence
Author Manuscript

grade 3 or higher there.

Additional inclusion criteria included: medial joint line tenderness on examination by an

experienced physical therapist (MJC) with tenderness over the patella less than medial
tenderness; stable medication regimen for three months; and willingness to wear insoles in
shoes at least 4 hours daily.

Exclusion Criteria—Subjects were excluded for the following reasons: a history of high
tibial osteotomy, other realignment surgery or a knee replacement in the painful knee; a knee
arthroscopy within the last 6 months or an intraarticular injection of either steroid or
viscosupplementation in that knee within the last 3 months. Persons with the following
disorders or conditions were excluded: rheumatoid arthritis or other inflammatory arthritis,
diabetic neuropathic pain or fibromyalgia; foot or ankle problems that contraindicated the
Author Manuscript

use of footwear load-modifying interventions; or severe coexisting medical morbidities.

Further exclusions included: inability to walk unaided without a crutch, cane or walker; BMI
>= 35; and persons who currently used or required foot orthoses. We also excluded those
unable to retain information regarding study procedures or who were unable to walk 100
meters without stopping. A secondary outcome of the study was change in MRI features,
and those with contraindications to MRI were also excluded, as were those who planned
knee surgery within the next 6 months.

Evaluation and Treatment

For eligible subjects, a gait laboratory appointment was made. There, subjects were
randomized remotely to the order of testing of insoles in their shoes (no insole, neutral
insole, lateral wedge insole). A 10 camera Qualisys (Gothenburg, Sweden) Proreflex motion
Author Manuscript

analysis system operating at 100 Hz and two Kistler (Alton, UK) force plates operating at
1000 Hz was used to measure kinematics and kinetics. Each participant completed a
minimum of five successful trials (a successful trial was when a single foot contacted the
force plate). The CAST marker set technique (9) was employed (8). Retroreflective markers
were mounted securely to the participants’ shoes with the foot modelled as a rigid segment.
Ankle and knee joint centers were calculated as midpoints between the malleoli and femoral
epicondyles respectively. The hip joint center was calculated using the regression model of
Bell et al (10) based on anterior and posterior superior iliac spine markers. Participants
walked at their self-selected speed and this did not differ across conditions. Using an inverse
dynamic approach Visual 3D (C-Motion, Rockville, Maryland), we calculated the first peak
KAM normalized to the participant’s mass (Nm/kg) averaged across the 5 trials. In addition,
subjects were asked about comfort of each condition and whether they noted any immediate
Author Manuscript

effect on knee pain.

We characterized a subject as a biomechanical responder if, for the study knee, there was at
least a 2% reduction of their KAM in the lateral wedge insole compared to both the KAM in
their own shoe and the KAM when they used the neutral insole. Biomechanical responders
were eligible for the study. We chose a 2% reduction as this is above the minimal detectable
difference in KAM (11) and was a reasonable approach to ensure biomechanical response

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 4

Eligible subjects were randomized by a statistician who had no contact with study staff and
Author Manuscript

created sealed opaque envelopes for each study ID number opened when a subject entered
the study. The randomized allocation list used to create the envelopes was a single-allocation
computer-generated list of balanced permuted blocks with a block size of 6. Study staff were
blinded to block size. Subjects were randomized to one of two treatment sequences 1:1 (AB
or BA) in a two-period crossover trial. Each randomized participant was provided either (A)
a 5° lateral wedge insole or (B) a neutral insole, for 8 weeks before an 8 week washout.
Then, they switched to the other treatment for 8 weeks. The initial treatment was taken away
at the last visit of the assigned treatment period, preventing treatment contamination. Both
insoles had a density of 70 Shore A.

Participants attended the clinic for 5 visits: a screening visit (−2 weeks), baseline visit (week
0), post-treatment visit (8 weeks), post-washout visit/second baseline visit (16 weeks), and
second post treatment visit (24 weeks).
Author Manuscript

Participants were asked to wear insoles at least 4 hours per day but could wear them for as
long as they wanted.

Bone marrow lesions (BMLs) appear on knee MRI in regions where excessive loading has
produced bone damage. Medial lesions predominate in those with medial OA. BMLs in the
patellofemoral joint shrank over 6 weeks when focal loading there was reduced (12),
suggesting BMLs may respond to unloading treatments. We evaluated subjects for change in
BML volume in the medial joint. At baseline, subjects underwent MRI of their study knee
and then obtained MRI’s again at 8, 16 and 24 weeks. Using a 3 T Philips Achieva scanner
(Philips, Best Netherlands), we obtained sagittal images with SPAIR fat suppression,
repetition time (TR) 4300ms, echo time (TE) 50ms, Field of View (FOV) 16×24cm,
Author Manuscript

212×220 acq matrix, slice thickness 3mm with 0.3mm gap BW 621–655 Hz/pixel.

Outcomes: Pain
The primary pre-specified symptom outcome was overall pain in the knee in the past week
scored using a numerical rating scale (0–10) as per recommendations of the IMMPACT
group (13). Also, at baseline, subjects were asked to identify the activity that provoked the
most knee pain and were asked to score pain during this activity at each visit) (14). Lastly,
we administered the KOOS questionnaire at each visit (15).

Outcomes : Structure
Technicians at iMorphics, blinded to treatment assignment, manually segmented BML
volumes in paired images from each subject’s knee. BMLs were outlined on each MRI slice
Author Manuscript

and the volume integrated over all slices. For sagittal images, on which we based results,
they segmented BMLs in patella, femur and tibia. Interobserver reliability for BML volume
was ICC = 0.91 (p <0.001) (12).

The primary structural outcome was change in medial BMLs. We defined these as BMLs
involving either the tibia medial to the cruciate ligaments or femur medial to the notch using
regions derived from the WORMS scale (16).

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 5

Analysis:
Author Manuscript

The analysis followed an intention-to-treat approach. Multiple imputation by chained

equations (MICE) was used to correct estimates for bias due to missing data, assuming data
was missing at random.

To assess the difference in treatment effects between the two insoles, we used maximum
likelihood mixed-effects multiple linear regression models. The primary analysis model used
overall pain in the last week, measured at the end of each treatment visit as the outcome and
adjusted for baseline (pre-treatment visit) pain scores. Using the baseline value as a
covariate (the ANCOVA approach) is a recommended methodology with less bias and
greater power to detect differences (17). Participant ID was included as a random effect. To
test for carry-over effects, we included three additional covariates: a treatment term, a period
term, and, the term testing for carryover, a treatment-by-period interaction term. We ran a
Author Manuscript

second model identical to the first that removed the period, and treatment-by-period
interaction terms.

The same regression models were used to analyze the secondary outcomes, using the same
terms, but replacing overall knee pain in the last week with the secondary outcome of
interest.

In addition to evaluating the response to insoles on a continuous scale, we also defined

responders as those who achieved an accepted minimally important difference in overall
knee pain of 1 on the 0–10 point scale (18). We examined what the percentage of trial
participants achieved responder status as the improvement in pain from baseline pain score
in that treatment period to the pain score at the end of the period.
Author Manuscript

Sample Size
We aimed to detect a treatment effect size (delta) of 0.4 SDs with 80% power (two sided
alpha = 0.05). Based on a within-person SD of 2.4 for overall pain in the last week (19)
testing a lateral wedge insole, this effect size represents a difference of 1.0 point on this knee
pain rating scale and translates to a sample size of 52 subjects completing the trial. We
assumed 10% loss to follow-up and aimed to randomize 58 subjects.

The protocol was approved by National Research Ethics Service Committee North West
(Preston).

RESULTS
Of 83 participants who satisfied inclusion criteria and were evaluated in the gait laboratory
Author Manuscript

(Figure 1), 62 were biomechanical responders. In this group, the mean reduction in KAM
compared with the neutral insole was 6.6% and compared with their own shoe was 7.5%
(Table 1). In contrast, 21 participants (25.3%) were biomechanical non-responders and, on
average, their KAM was 2% higher with the lateral wedge insoles than with their own shoes
(Table 1). There were no differences between responders and non-responders in immediate
knee pain on walking in the lab or in the comfort of their own shoes vs. inserts. Randomized
participants were on average 64 years old, had an average BMI of 28, and the majority were

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 6

men (Table 2). Of the 62 randomized, 59 completed the first and 56 the second treatment
Author Manuscript

period.

We found no significant evidence of carry-over effects for any of the outcomes of interest.
When we examined our primary outcome (Figure 2 and Table 3), we found that subjects
reported less knee pain when randomized to lateral wedge insoles than when on neutral
insoles (difference in pain score 0.7 on a 0–10 scale (95%CI 0.1, 1.2; p = 0.02). The pain
during their most painful nominated activity was also less during the period on lateral wedge
insoles (Table 3). However, we did not find significant differences for the KOOS pain score
or the other KOOS subscales.

We found no effect of treatment on medial bone marrow lesion volume or on total BML
volume in the knee (Table 3).
Author Manuscript

Results from complete case analyses were similar.

During their period on active treatment, 28% of participants (16/57) achieved a minimally
important improvement, and on neutral insoles, 22% (13/58) experienced this level of
improvement. The odds ratio for achieving important improvement on lateral wedges vs.
neutral inserts was 1.35 (95% 0.58, 3.13, p = 0.49).

At the end of the first treatment period, participants reported mean insole use of 7.10 hrs/day
(SD 2.72), and at the end of the second, participants reported use of 7.80 hrs/day (SD 3.17).

During the trial, 7 participants experienced side effects leading to temporary (3) or
permanent (4) treatment discontinuation. Of these, 4 occurred during lateral wedge and 3
during neutral insole treatment. Of the 2 who discontinued treatment on lateral wedge
Author Manuscript

insoles, one had calf pain at night and the other experienced worse knee pain. Of the 2
stopping neutral insoles, one developed a toe blister and the other had worsening knee pain.

DISCUSSION
In this trial, we prescreened subjects to select biomechanical responses to lateral wedge
insoles. When we excluded persons who had either patellofemoral involvement or failed to
show biomechanical responses to insoles, we detected a small effect of insoles on pain
reduction that was missed in previous trials in which there was no prescreening. However,
most participants in the trial did not experience a level of improvement that would qualify as
minimally important based on accepted thresholds. While this trial suggests an approach to
detect efficacy of this treatment, the efficacy was not sufficient to recommend this treatment
and the screening approach we used. In this case, we may have found a treatment that shows
Author Manuscript

statistical without clinical significance.

The small treatment effect we found may account for the lack of consistency across
outcomes. While the lateral wedge insole reduced knee pain based on our primary outcome,
knee pain over the past week and also led to a reduction in the patient’s nominated painful
activity, it did not significantly reduce pain or self-reported function assessed using the
KOOS survey, an expanded version of the Western Ontario McMaster Universities

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 7

Osteoarthritis Index (WOMAC). We have reported that WOMAC is not as sensitive to

Author Manuscript

change as global questions about knee pain (14) and that may partly explain the lack of
effect. However, this lack of effect also suggests that the treatment effect was modest. For
the primary outcome, the effect size calculated as a standardized response mean was 0.30.
This effect is comparable to that found for nonsteroidal drugs vs. acetaminophen in OA (20).
While a validated estimate of minimally important improvement (18) for our primary
outcome is 1.0, the mean difference in pain reduction between the lateral wedge and neutral
insole periods of treatment was only 0.7.

Our screening process utilized a gait laboratory to identify participants with a biomechanical
response to wedges. Such laboratory evaluations are expensive and may not be widely
available. Clinical screening protocols could be developed that might identify with high
probability those likely to respond. We have tried to develop such a protocol without success
(21), but other efforts are needed. While the KAM measures the medial vs. lateral load,
Author Manuscript

another factor affecting medial joint loading is the knee flexion moment. We have previously
reported that the lateral wedge used in this study does not affect this moment (22).

Once patellofemoral OA is ruled out by a simple physical examination, given the low cost
and benign safety profile of these wedges, it might be argued that treating patients with these
insoles is a reasonable clinical strategy rather than seeking a gait laboratory evaluation. Even
so, the treatment effect is likely to be small and only a few patients will note substantial
benefit.

Could the effect of lateral wedge insoles have been similar to that seen in previous trials that
used WOMAC or KOOS as outcomes? In two large studies using variable stiffness shoes or
lateral wedge insoles, authors (5, 19) used the same global knee pain measure and found no
Author Manuscript

effect of treatment vs. control on pain.

Osteoarthritis has been challenging to treat because it combines mechanopathology with an

inflammatory response to joint injury, both of which contribute to pain and disease
progression. It has been unclear whether nonsurgical treatments targeting pathomechanics
were likely to be major elements of the treatment armamentarium. Knee brace adherence is
poor, for example. Our study offers modest promise for a simple, inexpensive treatment.
Further refinement of the treatment with use of specific shoes or increases in the degree of
wedging may increase efficacy.

We and others have reported that biomechanical response to lateral wedge insoles is variable
(22, 24, 25). Although the reasons are unclear, one study (25) suggests that stiffness in feet
and ankles in some persons may prevent the lateral ankle eversion that is necessary for knee
Author Manuscript

loading to change with this treatment.

While lateral wedge insoles are thought to be safe treatments, their use occasionally (21)
generates complaints of discomfort in the shoe and may cause back pain (19). We did not
find major safety concerns, and no one discontinued treatment due to back pain.

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 8

We carried out a crossover trial to evaluate the effect of lateral wedge insoles. Crossover
Author Manuscript

trials permit the testing of treatments more efficiently than parallel design trials and make it
easier to detect modest effects of treatments.

In summary, we found, for the first time, that lateral wedge insoles may be modestly
effective in reducing pain in persons with medial knee OA. However, the treatment effect
was small and most treated patients did not achieve conventional levels of minimal
important response. Future modifications of the treatment or of the screening strategy might
offer greater levels of efficacy.

Acknowledgements:
We appreciate the participation of all subjects and also are indebted to Helen Williams and Michelle Reading for
their help with the study. We also thank Michael LaValley, Tuhina Neogi and Catherine Hill for their valuable
suggestions.
Author Manuscript

Supported by the NIHR Manchester Biomedical Research Centre. Dr. Felson was supported by NIH AR47785

References
1. Lawrence RC, Felson DT, Helmick CG, Arnold LM, Choi H, Deyo RA, et al. Estimates of the
prevalence of arthritis and other rheumatic conditions in the United States. Part II. Arthritis Rheum
2008;58(1):26–35. [PubMed: 18163497]
2. Kerrigan DC, Lelas JL, Goggins J, Merriman GJ, Kaplan RJ, Felson DT. Effectiveness of a lateral-
wedge insole on knee varus torque in patients with knee osteoarthritis. ArchPhysMedRehabil
2002;83(7):889–93.
3. Arnold JB, Wong DX, Jones RK, Hill CL, Thewlis D. Lateral Wedge Insoles for Reducing
Biomechanical Risk Factors for Medial Knee Osteoarthritis Progression: A Systematic Review and
Meta-Analysis. Arthritis Care Res (Hoboken) 2016;68(7):936–51. [PubMed: 26605535]
4. Parkes MJ, Maricar N, Lunt M, LaValley MP, Jones RK, Segal NA, et al. Lateral wedge insoles as a
Author Manuscript

conservative treatment for pain in patients with medial knee osteoarthritis: a meta-analysis. JAMA
2013;310(7):722–30. [PubMed: 23989797]
5. Hinman RS, Wrigley TV, Metcalf BR, Campbell PK, Paterson KL, Hunter DJ, et al. Unloading
Shoes for Self-management of Knee Osteoarthritis: A Randomized Trial. Ann Intern Med
2016;165(6):381–9. [PubMed: 27398991]
6. Butler RJ, Marchesi S, Royer T, Davis IS. The effect of a subject-specific amount of lateral wedge
on knee mechanics in patients with medial knee osteoarthritis. J Orthop Res 2007;25(9):1121–7.
[PubMed: 17469197]
7. Hinman RS, Bowles KA, Bennell KL. Laterally wedged insoles in knee osteoarthritis: do
biomechanical effects decline after one month of wear? BMCMusculoskeletDisord 2009;10:146.
8. Jones RK, Chapman GJ, Parkes MJ, Forsythe L, Felson DT. The effect of different types of insoles
or shoe modifications on medial loading of the knee in persons with medial knee osteoarthritis: a
randomized trial. J Orthop Res 2015;33(11):1646–54. [PubMed: 25991385]
9. Cappozzo A, Catani F, Croce UD, Leardini A. Position and orientation in space of bones during
Author Manuscript

movement: anatomical frame definition and determination. Clin Biomech (Bristol, Avon)
1995;10(4):171–8.
10. Bell AL, Brand RA, Pedersen DR Prediction of hip joint centre location from external landmarks.
Hum Mov Sci 1989;8(1):3–16.
11. Jones RK, Chapman GJ, Findlow AH, Forsythe L, Parkes MJ, Sultan J, et al. A new approach to
prevention of knee osteoarthritis: reducing medial load in the contralateral knee. J Rheumatol
2013;40(3):309–15. [PubMed: 23322462]

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 9

12. Callaghan MJ, Parkes MJ, Hutchinson CE, Gait AD, Forsythe LM, Marjanovic EJ, et al. A
randomized trial of a brace for patellofemoral osteoarthritis targeting knee pain and bone marrow
Author Manuscript

lesions. Ann Rheum Dis 2015;74(6):1164–70. [PubMed: 25596158]

13. Dworkin RH, Turk DC, Farrar JT, Haythornthwaite JA, Jensen MP, Katz NP, et al. Core outcome
measures for chronic pain clinical trials: IMMPACT recommendations. Pain 2005;113(1–2):9–19.
[PubMed: 15621359]
14. Parkes MJ, Callaghan MJ, O’Neill TW, Forsythe LM, Lunt M, Felson DT. Sensitivity to Change of
Patient-Preference Measures for Pain in Patients With Knee Osteoarthritis: Data From Two Trials.
Arthritis Care Res (Hoboken) 2016;68(9):1224–31. [PubMed: 26713415]
15. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee Injury and Osteoarthritis
Outcome Score (KOOS)--development of a self-administered outcome measure. JOrthopSports
PhysTher 1998;28(2):88–96.
16. Peterfy CG, Guermazi A, Zaim S, Tirman PF, Miaux Y, White D, et al. Whole-Organ Magnetic
Resonance Imaging Score (WORMS) of the knee in osteoarthritis. OsteoarthritisCartilage
2004;12(3):177–90.
17. Vickers AJ, Altman DG. Statistics notes: Analysing controlled trials with baseline and follow up
Author Manuscript

measurements. BMJ 2001;323(7321):1123–4. [PubMed: 11701584]

18. Salaffi F, Stancati A, Silvestri CA, Ciapetti A, Grassi W. Minimal clinically important changes in
chronic musculoskeletal pain intensity measured on a numerical rating scale. Eur J Pain 2004;8(4):
283–91. [PubMed: 15207508]
19. Bennell KL, Bowles KA, Payne C, Cicuttini F, Williamson E, Forbes A, et al. Lateral wedge
insoles for medial knee osteoarthritis: 12 month randomized controlled trial. BMJ
2011;342:d2912. [PubMed: 21593096]
20. Verkleij SP, Luijsterburg PA, Bohnen AM, Koes BW, Bierma-Zeinstra SM. NSAIDs vs
acetaminophen in knee and hip osteoarthritis: a systematic review regarding heterogeneity
influencing the outcomes. Osteoarthritis Cartilage 2011;19(8):921–9. [PubMed: 21619937]
21. Chapman GJ, Parkes MJ, Forsythe L, Felson DT, Jones RK. Ankle motion influences the external
knee adduction moment and may predict who will respond to lateral wedge insoles?: an ancillary
analysis from the SILK trial. Osteoarthritis Cartilage 2015;23(8):1316–22. [PubMed: 25749010]
22. Jones RK, Chapman GJ, Forsythe L, Parkes MJ, Felson DT. The relationship between reductions in
knee loading and immediate pain response whilst wearing lateral wedged insoles in knee
Author Manuscript

osteoarthritis. J Orthop Res 2014;32(9):1147–54. [PubMed: 24903067]

23. Kris MG, Johnson BE, Berry LD, Kwiatkowski DJ, Iafrate AJ, Wistuba, II, et al. Using
multiplexed assays of oncogenic drivers in lung cancers to select targeted drugs. JAMA
2014;311(19):1998–2006. [PubMed: 24846037]
24. Hinman RS, Bowles KA, Payne C, Bennell KL. Effect of length on laterally-wedged insoles in
knee osteoarthritis. Arthritis Rheum 2008;59(1):144–7. [PubMed: 18163399]
25. Lewinson RT, Vallerand IA, Collins KH, Wiley JP, Lun VMY, Patel C, et al. Reduced knee
adduction moments for management of knee osteoarthritis:: A three month phase I/II randomized
controlled trial. Gait Posture 2016;50:60–8. [PubMed: 27580080]
Author Manuscript

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 10

Key Messages:
Author Manuscript

1. Trials testing lateral wedge vs. neutral insoles in persons with painful medial
knee osteoarthritis have shown no reduction in knee pain despite a proven
reduction in medial loading.

2. The biomechanical response to lateral wedges is inconsistent with roughly

25% of persons showing no reduction in medial loading, and shifting load
laterally may worsen patellofemoral loading.

3. This trial prescreened subjects with painful medial OA to remove those with
patellofemoral disease and those who did not show a reduction in medial knee
loading with lateral insoles.

4. While the trial showed a reduction in the primary pain outcome, this effect
Author Manuscript

was small and other measures of pain showed no effect, nor was there a
reduction in local bone marrow lesions on MRI, a measure of medial load.

5. Even after prescreening, lateral wedge insoles, at least as conventionally used,

do not produce important improvement in pain in most persons.
Author Manuscript
Author Manuscript

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 11
Author Manuscript
Author Manuscript
Author Manuscript

Figure 1.
Author Manuscript

Trial patient Flow Diagram

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Felson et al. Page 12
Author Manuscript
Author Manuscript
Author Manuscript

Figure 2.
Cross-Over Change in Overall Knee Pain in Last Week (NRS)
Author Manuscript

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 1.

Percent difference in responders vs. non-responders in external Knee Adduction Moment (KAM) using lateral wedge insoles vs. neutral insoles and vs.
participant’s own shoes
Felson et al.

Neutral
Own Shoe Lateral Wedge Insole Difference Between Own Shoe and LWI Difference Between Neutral Insole and LWI
Insole
Adjusted mean adjusted mean
mean (SD) mean (SD) mean (SD)
(95% CI) (95% CI)
Absolute Values 0.54 (0.17) 0.54 (0.16) 0.55 (0.17) 0.01 (0.00 to 0.02) 0.00 (−0.01 to 0.02)
% Change - - - 1.98 (−0.07 to 4.03) 0.64 (−2.12 to 3.39)
Non-Responders (N=21)
Immediate Pain 2.86 (2.46) 2.33 (2.35) 2.05 (2.09) −0.78 (−1.28 to −0.28) −0.27 (−0.77 to 0.23)
Immediate Comfort* 7.29 (1.74) 7.10 (2.21) 7.62 (1.77) 0.35 (−0.18 to 0.88) 0.22 (−0.32 to 0.75)
Absolute Values 0.50 (0.15) 0.49 (0.15) 0.46 (0.14) -0.04 (−0.04 to −0.03) -0.03 (−0.04 to −0.03)
% Change - - - −7.54 (−8.53 to −6.55) −6.56 (−7.69 to −5.42)
Responders (N=62)
Immediate Pain 3.03 (2.21) 2.84 (2.06) 2.48 (1.89) −0.28 (−0.57 to 0.02) −0.18 (−0.48 to 0.11)
Immediate Comfort* 7.00 (1.87) 7.32 (1.94) 7.44 (1.77) 0.14 (−0.21 to 0.49) 0.12 (−0.22 to 0.47)

Immediate pain and comfort graded on a 0–10 scale but for pain, lower numbers represent less pain and for comfort, higher scores represent more comfort.

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

Page 13
 Felson et al. Page 14

Table 2.

Characteristics of responders and non-responders

Author Manuscript

Responders
Variable (N=62) Non-Responders (N=21)
Age in years mean (SD) 64.18 (9.10) 65.86 (10.03)
BMI mean (SD) 28.21 (3.44) 28.56 (3.99)
No. females (%) 23 (37.10) 9 (42.86)

HAD* Anxiety Score mean (SD) 12.17 (2.24) 12.48 (1.57)

HAD* Depression Score mean (SD) 9.10 (1.24) 8.33 (0.97)

Overall Knee Pain in last week NRS (0–10) mean (SD) 5.26 (1.63) 5.24 (1.87)
Pain on nominated activity NRS (0–10) mean (SD) 6.18 (1.54) 6.05 (1.66)

KOOS Pain Subscale Score (0–100)ɸ mean (SD) 55.20 (13.45) 58.07 (12.07)

Kellgren and Lawrence Grade of

Studied Knee (%)
Author Manuscript

-Grade 2 17 (27.4%)
-Grade 3 37 (59.7%)
-Grade 4 8 (12.9%)

*
Hospital Anxiety and Depression Scale. Scores range from 0–21 with higher scores (above 11) indicating either depression or anxiety
ɸ
KOOS scores range from 100 to 0 where 100 represents no pain/difficulty.
Author Manuscript
Author Manuscript

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 3.

Comparison between lateral wedge and neutral insoles after 8 weeks’ treatment**

Post-Treatment Adjusted Mean (95% CI)

Difference between treatments
Felson et al.

Outcome Lateral Wedge Insole Neutral Insole (mean, 95% CI, p)

Pain in last week* 4.16 (3.69 to 4.62) 4.85 (4.42 to 5.28) 0.70 (0.12 to 1.27), 0.02

Pain in nominated activity* 4.80 (4.30 to 5.31) 5.77 (5.28 to 6.26) 0.97 (0.32 to 1.61), 0.003

KOOS Pain subscale score* 60.66 (57.21 to 64.11) 58.82 (55.67 to 61.96) −1.84 (−6.31 to 2.62), 0.42

KOOS Symptom subscale score 60.64 (57.59 to 63.70) 59.41 (56.40 to 62.42) −1.23 (−5.11 to 2.65), 0.53
KOOS Activities of Daily Living subscale score 66.29 (63.15 to 69.44) 65.01 (61.88 to 68.14) −1.28 (−5.19 to 2.62), 0.52
KOOS Sports & Recreation subscale score 43.57 (39.46 to 47.67) 42.21 (37.56 to 46.86) −1.36 (−6.97 to 4.26), 0.63
KOOS Quality of Life subscale score 44.18 (40.62 to 47.73) 44.09 (40.80 to 47.38) −0.09 (−4.64 to 4.47), 0.97

Total BML Volume 12959.66 (10991.04 to 14928.29) 11047.29 (8833.52 to 13261.05) −1912.38 (−4602.61 to 777.86), 0.16
Medial TF BML Volume 8331.48 (6903.42 to 9759.55) 7051.42 (5398.37 to 8704.47) −1280.07 (−3210.91 to 650.78), 0.19
Lateral TF BML Volume 1340.93 (801.87 to 1879.99) 1219.47 (648.63 to 1790.32) −121.45 (−630.27 to 387.36), 0.64

*
For pain in last week and pain in nominated activity, lower scores represent pain reduction. For KOOS higher scores represent pain reduction.
**
Pain and other scores in the table are adjusted for the baseline value of that outcome which is the same for both treatment groups.

Arthritis Rheumatol. Author manuscript; available in PMC 2020 June 01.

Page 15