Gait & Posture 91 (2022) 137–148

Contents lists available at ScienceDirect

Gait & Posture

journal homepage: www.elsevier.com/locate/gaitpost

Review

Treatment of knee hyperextension in post-stroke gait. A systematic review

Marieke Geerars a, b, *, Nympha Minnaar-van der Feen a, Bionka M.A Huisstede a, c
a
Physical Therapy Sciences, Program in Clinical Health Sciences, University Medical Center Utrecht, Heidelberglaan 100, 3584 CX Utrecht, the Netherlands
b
Stichting Axioncontinu, Rehabilitation Center de Parkgraaf, Beneluxlaan 926, 3526KJ Utrecht, the Netherlands
c
Department of Rehabilitation, Physical Therapy Science and Sports, Brain Center, University Medical Center Utrecht, Utrecht University, Heidelberglaan 100, 3584 CX
Utrecht, the Netherlands

A R T I C L E I N F O A B S T R A C T

Keywords: Background: Post-stroke, patients exhibit considerable variations in gait patterns. One of the variations that can
Genu recurvatum be present in post-stroke gait is knee hyperextension in the stance phase.
Knee hyperextension Research question: What is the current evidence for the effectiveness of the treatment of knee hyperextension in
Stroke
post-stroke gait?
Hemiparesis
Gait
Methods: MEDLINE, EMBASE, PEDro, CINAHL, and the Cochrane library were searched for relevant controlled
trials. Two researchers independently extracted the data and assessed the methological quality. A best evidence
synthesis was conducted to summarize the results.
Results: Eight controlled trials (5 RCTs, 3 CCTs) were included. Three types of interventions were identified:
proprioceptive training, orthotic treatment, and functional electrostimulation (FES). In the included studies, the
time since the stroke occurrence varied from the (sub)acute phase to the chronic phase. Only short-term effects
were investigated. The adjustment from a form of proprioceptive training to physiotherapy training programs
seems to be effective (moderate evidence) for treating knee hyperextension in gait, as applied in the subacute
phase post-stroke. Neither evidence for effects on gait speed nor gait symmetry were found as a result of pro­
prioceptive training. Orthoses that cover the knee have some effects (limited evidence) on knee hyperextension
and gait speed. No evidence was found for FES.
Significance: This is the first systematic literature review on the effectiveness of interventions on knee hyper­
extension in post-stroke gait. We found promising results (moderate evidence) for some “proprioceptive ap­
proaches” as an add-on therapy to physiotherapy training programs for treating knee hyperextension during the
subacute phase post-stroke, in the short-term. Therefore, initially, clinicians should implement a training pro­
gram with a proprioceptive approach in order to restore knee control in these patients. Because only studies
reporting short-term results were found, more high-quality RCTs and CCTs are needed that also study mid- and
long-term effects.

1. Introduction depending on the residual function and the severity of sensorimotor

impairment [5]. One of the variations that can be present in post-stroke
Stroke is a global health problem. In 2016, approximately 13.7 gait is hyperextension of the knee in the paretic limb during the stance
million strokes occurred worldwide. There are estimated to be 80.1 phase [6]. Knee hyperextension is also known as genu recurvatum,
million stroke survivors globally [1]. which is characterized by a ground reaction force vector, which passes
Over the past two decades, the number of stroke survivors has in front of the knee resulting in full knee extension (0◦ ) or more [7].
increased [2]. For everyone who suffers a stroke, rehabilitation is an Knee hyperextension is a progressive, disabling deformity that impairs
important aspect of recovery [3]. Of all the possible rehabilitation goals, walking speed, decreases gait efficiency, increases use of energy during
being able to walk again is high on the list for most stroke survivors [4]. walking, and can result in or might be associated with knee pain [8–10].
Post-stroke, patients exhibit considerable variations in gait patterns, Moreover, knee hyperextension can reduce gait symmetry and thus,

* Corresponding author. Present address: Beneluxlaan 926 3526KJ Utrecht, the Netherlands.
E-mail addresses: [email protected] (M. Geerars), [email protected] (N. Minnaar-van der Feen), [email protected]
(B.M.A Huisstede).

https://doi.org/10.1016/j.gaitpost.2021.08.016
Received 23 October 2020; Received in revised form 6 August 2021; Accepted 22 August 2021
Available online 24 August 2021
0966-6362/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
M. Geerars et al. Gait & Posture 91 (2022) 137–148

affect cosmetic appearance. The asymmetric gait pattern generally re­ 2.4. Methodological quality
sults in the increased duration of the swing phase of the paretic limb
[11]. Post-stroke, 20–68 % of the walking population reports hyperex­ Two reviewers (MG and NM) independently assessed the methodo­
tension of the knee [7,8,12,13]. The knee hyperextension can be caused logical quality of the included trials. A consensus method was used in the
by either a single symptom or a combination of symptoms such as cases where there were any disagreements.
weakness, spasticity or retraction of the paretic limb muscles, limited The Downs and Black checklist [18] was used to assess the meth­
ankle mobility, proprioceptive disorders, and diminished velocity odological quality of the included trials (Appendix 2). This checklist is
properties of the distal limb muscles [14]. considered relevant for both randomized and nonrandomized studies. It
In scientific literature, several interventions have been reported for consists of 27 questions. Each question was scored as “yes” (1 or 2), “no”
the treatment of knee hyperextension in post-stroke gait. In 2010, (0) or “unable to determine” [19–21].
Bleyenheuft et al. [7] conducted a systematic review that studied in­ Two questions were modified: Question 15 (blinding measurement
terventions to treat knee hyperextension in hemiparetic adults. They of main outcome) was also given a score of 1 if measurements were done
reported on the effectiveness of retraining methods (functional elec­ with cameras, graphic recorders or potentiometers, which were
trostimulation (FES) and electro goniometric feedback), surgical treat­ considered to minimize the risk of bias. In this way, we could make a
ment, and orthotic treatment. distinction between trials that measured the occurrence or grade of knee
Since 2010, studies reporting on other treatments such as proprio­ hyperextension by personal perception or using an objective measuring
ceptive training and interventions to influence spasticity in the lower tool. Question 27 (power/sample-size calculation) was modified from a
limb muscles on knee hyperextension post-stroke have also been pub­ 5-point score to a 0/1-point score as was done in other studies [19,20,22,
lished [12,15,16]. However, until now, healthcare professionals have 23]: A score of 1 was given if a power or sample-size calculation was
not had a clear treatment preference. Therefore, the aim of the current reported and a 0 was given where there was no power or sample-size
study was to systematically review the scientific literature in order to calculation or a study lacked an explanation as to whether the
provide more clarity on the preferred method to address knee hyper­ sample-size was appropriate to detect a clinically important effect. Each
extension in post-stroke gait. paper was considered to have an “excellent” (24–28 points), “good”
(19–23 points), “fair” (14–18 points) or “poor” (<14 points) methodo­
2. Methods logical quality.

This systematic review was performed according to the Preferred 2.5. Data extraction
Reporting Items for Systematic review and Meta-Analysis Protocols
(PRISMA-P) 2015 statement [17]. Two reviewers (MG and NM) independently extracted the data, using
a data extraction form. Disagreements were addressed by means of
2.1. Literature search discussions. Information was collected on study design, study popula­
tion, aetiology of knee hyperextension, interventions, and outcome
A computer-aided literature search was carried out in MEDLINE, measures. Follow-up periods were categorized as short-term (0–3
EMBASE, PEDro, CINAHL, and the Cochrane library up to May 2021. months), mid-term (4–6 months), and long-term (>6 months).
The search strategy, which was composed with the assistance of an
experienced librarian, is shown in Appendix 1. In addition, references of 2.6. Data Analysis/Synthesis
the included studies were checked by hand to identify additional rele­
vant studies. We considered pooling the results of the included trials in case pa­
tient characteristics, interventions, and outcome measures were similar.
2.2. Inclusion criteria In case a quantitative analysis was not possible, the results were sum­
marized by using a best-evidence synthesis. A trial was included in the
Trials carried out following a comparative study design, where a test best-evidence synthesis if a level of significance was reported. The rating
group and a control group were compared, were used in this current system was based on van Tulder et al. [24] and adapted from Hom­
review. This included control groups consisting of people post-stroke bergen et al. [25] (Table 1).
such as randomized controlled trials (RCT) and clinical controlled tri­
als (CCT), including cross-over and paired sample designs. Studies were
considered eligible for inclusion if they met all of the following criteria:
(1) the study population consisted of stroke survivors with knee hy­
perextension in gait, (2) the studies had a minimum of 10 participants, Table 1
Levels of evidence for differences between the groups.
(3) the studies reported on the effect of interventions in the treatment of
knee hyperextension, (4) the outcome measures that were assessed • Strong evidence: consistently significant findings* among multiple high-quality
included: measurements on knee hyperextension in stance phase of gait, RCT’s or other controlled intervention studies (rated “excellent” or “good” on the
Downs and Black checklist18)
gait velocity, gait symmetry, or energy cost, and (5) the full texts of the • Moderate evidence: consistently significant findings* among multiple low quality
articles were available, in Roman alphabet. Studies were included when RCTs or other controlled intervention studies, (rated “fair” or “low” on the Downs
at least 75 % of the participants had been diagnosed with a stroke and and Black checklist) and/or one high quality RCT or other controlled intervention
the results of these participants were reported separately. study (rated “excellent” or “good” on the Downs and Black checklist)
• Limited evidence: significant findings* within one low-quality RCT or other
controlled intervention study (rated “fair” or “low” on the Downs and Black
2.3. Study selection checklist)
• Conflicting evidence: provided by conflicting (significant) findings in RCTs or other
Two reviewers (MG and NM) independently applied the inclusion controlled intervention studies (<75 % of the studies reported consistent findings)
• No evidence: no significant findings were found
criteria to select relevant trials based on titles, abstracts, and full texts
respectively. A consensus method was used in the cases where there was *
The results were consistent when at least 75 % of the studies showed results
a disagreement. A third reviewer (BH) was consulted if no consensus in the same direction, which was defined according to significance (P < 0.05).
could be reached.

138
M. Geerars et al. Gait & Posture 91 (2022) 137–148

3. Results 3.4. Effectiveness of interventions

3.1. Characteristics of the included studies Various treatment options for knee hyperextension are available. The
interventions studied in the included trials could be divided in three
The computerized search distinguished 1545 potentially eligible categories: 1) proprioceptive training [15,28,29,31,32], 2) orthotic
studies. Of these 1506 were excluded based on titles and abstract. Two treatment [6,30,33] and 3) functional electro stimulation (FES) [33].
articles [26,27] were excluded because full text could not be obtained, Aman and colleagues [34] classified “active movement/balance training,
neither through contacting the authors of otherwise. Another 29 studies passive movement training, somatosensory stimulation training, somatosen­
were excluded after screening the full texts. Finally, a total of eight trials sory discrimination training, and combined/multiple system training” as
were included in this review: five RCTs [15,28–31] and three (non-­ proprioceptive interventions. In case the intervention of a study
randomized) CCTs [6,32,33] (Fig. 1). included in this systematic review matched one of these categories as
established by Aman et al. [34], we called it proprioceptive training.
Due to the heterogeneity of the included trials and the differences in the
3.2. Data extraction content of the treatments, statistical pooling of the data was impossible.
Therefore, we used a best-evidence synthesis (BES) to summarize the
Characteristics of the included studies are presented in Table 2. results. The evidence for the effectiveness of the interventions of the
included trials is presented in Table 4.

3.3. Methodological quality 3.4.1. Proprioceptive training

Five trials [15,28,29,31,32], in which the interventions were clas­
Table 3 presents the results of the quality assessment. Three trials sified as proprioceptive training according to the classification of Aman
[15,28,29] were determined to be of excellent quality, two trials [31,33] et al. [34], investigated the influence of retraining knee control in the
of good quality, and the others of fair quality. The most prevalent stance phase of gait by proprioceptive training. In the proprioceptive
methodological flaws were not mentioning adverse effects (100 %), not approach, participants were encouraged to avoid hyperextension by
blinding the participants to the intervention (88 %), and not providing a keeping the paretic knee in a slightly flexed position during gait, by
power calculation to detect a clinically important effect (88 %).

Fig. 1. Flowchart of the literature search.

# Studies were included when at least 75 % of the participants had been diagnosed with a stroke and the results of these participants were reported separately.

139
M. Geerars et al. Gait & Posture 91 (2022) 137–148

Table 2
Characteristics and outcomes of the selected studies.
Authors (year) Study population Etiology knee Intervention Control/comparison Outcome measures Results
Design, total hyper extension
number of patients,
phase post-stroke

Proprioceptive
training

Guo et al. [28] Stroke survivors Not specified Whole body vibration Same program with Number of knee Between groups:
(2015) Age 47–61y. training (30–60 s of vibration switched hyperextension (number of)
Pilot RCT Time since stroke vibration and 10 s rest off. occurrence in 10 m d = 1.749*, 95 %CI
n = 30 (days) (mean (SD)) intervals, 10 rounds per set, walking, at 8 weeks [2.915, 7.285]
subacute phase (control/ 8 sets per day) while follow up p < 0.001*.
post-stroke experimental group) standing on a platform in in favor of the
59.4(61.4)/66.9 semi-squatting position, experimental group.
(42.9) with even weight Within groups:
Participants had distribution and single leg (number of)
standing balance standing with knee bending before treatment/after
score ≥ 2. within 0–15◦ , for 8 weeks. treatment
In addition: regular Mean(SD)
exercises: Range of motion Control group 26.2
exercises of lower limb, PNF (5.4)/10.5(3.2)*
exercise, climbing stairs, p < 0.05
walking with brace and Experimental group
electrical stimulation. 24.8(4.7)/5.4(2.6)*, p
< 0.05*.
n = 15 n = 15 Maximum gait speed Between groups:
over 10 m.(sec), (sec)
At 8 weeks follow up d = 1.345, 95 %CI
[1.896, 6.704],
p = 0.001* in favor of
the experimental
group.
Within groups:
Mean(SD) (sec)
Control group 24.6
(4.5)/16.3(3.7)*
p < 0.05*
Experimental group
24.7(5.5)/12.0(2.6)*,
p < 0.05*.

Dalal et al. [15] Stroke survivors Not specified Prowling along with 6 sessions routine Mean knee Between groups:
(2018) Age 40–80y. proprioceptive training (i.e. physiotherapy for hyperextension Mean change (Median
RCT Time since stroke: n.r squats, single limb stance/ 45–60 min in 6–10 (degrees) change score (IQR)) knee
n = 32 Participants were able squats) for 15–20 min as an consecutive days. 0–6 sessions hyperextension
acute and to walk 5 m with or adjunct to 6 sessions of degrees
subacute phase without assistance or routine physiotherapy for Control vs
post-stroke a walking aid. 45–60 min in 6–10 experimental group
consecutive days. − 0.35(− 1.18, 1.75) vs
− 4.52(− 5.72, − 2.82) p
< 0.001*.
n = 16 n = 16 Within groups:
before treatment/after
treatment (Median
(IQR)) degrees
Control group:
8(5.25, 10.29)/7.87
(5.12, 9.96)
NS
Experimental group
8.67(7.17,10.40)/3.87
(3.26, 5.00)
p = 0.001*
Gait speed over 5 m Between groups:
(sec) Mean change Median
change score 0–6 (IQR)
sessions Control vs
experimental group
− 2.5(− 8.00, 0.00) vs
− 5.00(− 8.00, − 0.50)
NS
Within groups:
Mean change Median
(IQR) sec before/after
treatment,
Control group:

(continued on next page)

140
M. Geerars et al. Gait & Posture 91 (2022) 137–148

Table 2 (continued )
Authors (year) Study population Etiology knee Intervention Control/comparison Outcome measures Results
Design, total hyper extension
number of patients,
phase post-stroke

26(16.75, 31.50)/22.5
(13.75, 28.75)
p = 0.041*
Experimental group:
23(15, 29.75)/18
(11.25, 22.50), p =
0.002*.

Morris et al. [29] Stroke survivors Not specified In phase 1 (4 weeks): Conventional Peak Knee Extension Mean reduction PKE
(1992) Age (y)(Mean(SD)) Conventional therapy with therapy in both (PKE) i.e. maximum After Phase1:
RCT 64.2(11.9) electro goniometric phases (8 weeks) 5 angle of knee extension Between groups
n = 26 Time since stroke feedback during standing times per week for during stance phase Mann-Whitney test p =
subacute phase (days) (mean (SD)) 62 and gait training. (auditory (mean(SD) 27.0 (degrees); Change score 0.173 NS
post-stroke (37) signal at a certain knee (11.6) min. 0-4-8 weeks Mean reduction (SD)
Participants were able angle). degrees
to walk safe for 4 × 10 In phase 2 (4 weeks): Control vs
m without aids or conventional therapy experimental group
orthoses. without electro goniometric 2.4(4.2) vs 3.1(3.0)
feedback. After Phase2:
Treatment duration in both Between groups
phases: 5 times per week for Mann-Whitney test, p
(mean(sd) 27.0(11.6) min. = 0.011*
The conventional therapy Mean reduction (SD)
was a Motor Relearning degrees
Programme: gait training in Control vs
which stereotyped and experimental group
abnormal synergistic 0,4(3.1) vs 1.7(1.8)
activity was discouraged.
n = 13 n = 13 Gait symmetry After Phase1 the
measured by single limb symmetry improved:
support. Between groups:
Change score 0-4-8 Mean change
weeks Asymmetry (SD)
% of gait cycle
Control vs
experimental group:
6.1(11.9) % vs 2.7
(11.7) %
NS
After Phase2 the
symmetry regressed:
Between groups:
Mean change
Asymmetry (SD)
% of gait cycle
Control vs
experimental group:
2.4(12.1) % vs 0.4
(16.5) %
NS
Gait speed (m/min), at Mean speed
start, improvement improvement (SD) (m/
after phase1 and min):
phase2. After Phase1
Change score 0-4-8 Between groups:
weeks Control group vs
experimental:
0.8 (12.5) m/min vs
6.3 (10.2) m/min
NS
After Phase2:
Between groups:
Control group vs
experimental
5.8(11.2) m/min vs 8.2
(9.9) m/min
NS

Lee et al. [31] Stroke survivors Not specified Treadmill gait with assistive Conventional Knee hyperextension in Mean knee
(2017) Age y(Mean(SD) 39.2 guidance force using tubing treadmill gait (CTG) midstance, (degrees) hyperextension(SD)
Cross-over RCT (16.8), (GTG) to improve knee measured during a set degrees
n = 18 Time since stroke stabilization during of 5 min CTG and 5 min CTG(control) vs GTG
(and 15 healthy (months) (mean (SD)) midstance. GTG. (experimental) in the
(continued on next page)

141
M. Geerars et al. Gait & Posture 91 (2022) 137–148

Table 2 (continued )
Authors (year) Study population Etiology knee Intervention Control/comparison Outcome measures Results
Design, total hyper extension
number of patients,
phase post-stroke

adults, not 6.7(2.9) 2–5 sessions of 10 min CTG hemiparetic group:

included in this Participants had a and 10 min GTG were 189(3.9) vs 178(7.7)*
review) Berg Balance Score conducted for each p < 0.01*
subacute to >40. individual for modification
chronic phase of the tubing tension.
post-stroke
n = 18 n = 18

Ceceli et al. [32] Stroke survivors Not specified Joint-position biofeedback conventional Number of steps with Between groups:
(1996) Age experimental training (acoustic signal physiotherapy (i.e., knee hyperextension in significant difference
CCT group 13–67y when knee extension exercises for pelvis, 50 steps after 10 days of
n = 41 (median 49.5), age exceeded 180◦ (+/− 3◦ ) for hip control and In advance, after 10 training in favor of the
acute to chronic control group 12–65y 30 min per day, for 10 days weight shifting) days of training and experimental group, p
phase post-stroke (median 54). in addition to conventional after 6 months follow < 0.05*
Time since stroke 0- physiotherapy. up Within groups: Mean
>1y. number of steps with
Participants had a knee hyperextension
normal passive range before/after training
of motion in lower Control group: 48.67/
extremity, were able 39.38.
to walk unassisted for Experimental group
50 steps. 45.19/7.538.
After 6 months: In
control group 2/9
walked without knee
hyperextension vs 4/11
in the experimental
group.
No statistical analysis
available on six
months.
n = 26 n = 15
Gait speed (number of Gait speed: NS
steps per minute) in
advance and after 10
days of training.

Orthotic treatment

Portnoy et al. [30] Stroke survivors spasticity, paresis Use of a hinged soft knee No intervention Peak Knee Extension in Mean PKE (SD) degrees
(2015) Age (y) (mean(SD)): lower limb orthosis, set on 10◦ of knee stance phase (PKE) without orthosis vs
Cross-over RCT 59.9 (15,1) muscles flexion, for 4 weeks. Two (degrees) with orthosis:
n = 31 Time since stroke (y) groups A and B. Mean score at baseline − 8.2(7.2) vs 10.1(9.5)
subacute to (mean (SD)) 6.1(6,7), After initial measurement and after 4 weeks p < 0.001*
chronic phase range 3 months-25y Group A used orthosis for 4 wearing orthosis.
post-stroke Participants walked weeks, group B no Gait speed over 10 m
independently with or intervention. After second (sec)
without walking aid. measurement group B used Mean score at baseline Mean time over 10 m
orthosis for 4 weeks and and after 4 weeks (SD) sec. without
group A no intervention. wearing orthosis. orthosis vs with
Measurement 3. orthosis 22.9(16.6) vs
n = 31 n = 31 21.3(16.6), p = 0.011*
(n = 17 group A, n = 14 (n = 17 group A, n = Gait symmetry indices NS
group B) 14 group B) without and with knee
orthosis. Mean score at
baseline and after 4
weeks wearing orthosis.

Boudarham et al. Stroke survivors spasticity Use of a Knee-ankle-foot No intervention Peak Knee Extension in Mean PKE(SD) degrees
[6] (2012) Age (y) (mean (SD)): quadriceps (n = orthosis (KAFO) n = 11 stance phase (degrees) without KAFO vs with
CCT (Cross-over 51 (15). 6), spasticity of Two gait conditions were Score in one session KAFO
trial) Time since stroke m. triceps surae measured in a fixed order: − 16.2 (11.9) vs − 7.6
n = 11 72− 672 months. (n = 3), weakness without KAFO and with (7.4) p = 0.029*.
chronic phase Participants were able of quadriceps (n KAFO Gait speed over 10 m Mean speed(SD) m/sec
post-stroke to walk 10 m without = 2) n = 11 (m/sec) without KAFO vs with
walking aids. Score in one session KAFO
0.57(0.36) vs 0.73
(0.34), p = 0.025*,
Swing phase asymmetry Mean swing phase
duration ratio asymmetry duration
Score in one session ratio(SD) without
KAFO vs with KAFO:
1.93(0.77) vs 1.27
(0.10), p = 0.014*

(continued on next page)

142
M. Geerars et al. Gait & Posture 91 (2022) 137–148

Table 2 (continued )
Authors (year) Study population Etiology knee Intervention Control/comparison Outcome measures Results
Design, total hyper extension
number of patients,
phase post-stroke

(meaning symmetry
significantly increased)
Stance phase Mean stance phase
asymmetry duration asymmetry duration
ratio; ratio(SD) without
Score in one session KAFO vs with KAFO
0.82(0.11) vs 0.86
(0.08), NS
Step length (non) Mean step length(SD)
paretic limb without m. without KAFO vs
and with KAFO(m) with KAFO: non-
Score in one session paretic limb
0.35(0.18) vs 0.40
(0.20)
p < 0.05*;
paretic limb:
0.42(0.16) vs 0.48
(0.15), NS.

FES

Bae et al. [33] Stroke survivors Excessive plantar Comparison of three Barefoot walking Knee angle at mid Between groups: N.S.
(2019) Age (y) (mean(SD)): flexion ankle due walking conditions: and AFO walking stance (degrees) Within groups:
CCT (Cross-over 54.41(19.29), to weakness consecutively barefoot, Score in one session Mean angle (SD)
trial) Time since stroke dorsiflexor or AFO and FES degrees
n = 12 (months) (mean(SD)): spasticity plantar FES Barefoot 8.40(11.22)
chronic phase 13.16(9.73) flexor Foot drop stimulator on FES 9.53(11.66)
post-stroke peroneal nerve (back of AFO 8.20(11.25)
fibula head) and anterior
tibial muscle (5 cm under
fibula head)
n = 12 n = 12 Gait speed over 6 m Gait speed: N.S.
AFO walking Barefoot walking (cm/sec)
AFO featuring 5◦ of ankle Score in one session
dorsiflexion
n = 12 n = 12
Each walking modality was
repeated three times (six
meters, three minutes break
between each time).
Between each modality ten
minutes break was
mandatory.

Abbreviations: RCT: randomized controlled trial; CCT: controlled clinical trial; NS: Not Significant; PKE: Peak Knee Extension; FES: functional electrical stimulation;
AFO: Ankle foot orthosis; *: statistically significant; m: meters; CI: Confidence Interval; SD: Standard Deviation; d:difference; IQR: Inter Quartile Range; n:number.

instruction or biofeedback, and the quadricep muscles were trained in a program on its own for the effectiveness on knee hyperextension and
flexed position. gait speed in the short-term.

3.4.1.1. Adjustment of Whole-Body Vibration to a training program versus 3.4.1.2. Prowling and proprioceptive training adjusted to routine physio­
a training program. Guo et al. [28] (RCT, n = 30, excellent quality) therapy versus routine physiotherapy. Dalal et al. [15] (RCT, n = 32,
included participants who were 0–3 months post-stroke with a standing excellent quality) included persons in the acute and subacute phases
balance score ≥ 2.The eight-week training program contained exercises post-stroke. The participants had a Brunnstrom Recovery Stage (BRS) of
in a semi-squatted position, with even weight distribution, and single leg the affected lower extremity ≥3 (i.e. synergy dependent voluntary
standing with knee flexion within 0–15◦ as well as regular exercises. movement). Both the experimental and control groups received 6 ses­
During the training, participants were positioned on the platform of a sions of routine physiotherapy lasting 45–60 min for 6–10 consecutive
vibration machine. The experimental group received whole body vi­ days. As an adjunct, the experimental group needed to perform 15–20
bration during the training. The control group received the same min of prowling (walking with bilateral knee flexion with trunk in mild
training on the same platform, but the vibration machine was turned off. forward flexion) and proprioceptive training of the affected limb (partial
Short-term training effects (after 8 weeks) were measured by counting squats, single limb stance, single limb partial dynamic squats). In
the times that the knee hyperextended during a 10 m-walk and addition, the participants of the experimental group were instructed to
measuring gait speed. Both groups reported significant benefits based on walk with their knees flexed during routine daily activities. The content
the times that the knee hyperextended during the 10 m-walk. The of the routine physiotherapy was not described. Short-term training
experimental group improved significantly more based on knee hyper­ effects (after 6 sessions) were measured on the degree of knee hyper­
extension (degrees) as compared to the control group: 1.749, (2.951, extension and gait speed over a five-meter distance. Group comparisons
7.285), (mean difference, (95 % CI)), p < 0.001 and gait speed in m/s: showed significant benefit to knee hyperextension (degrees) in favour of
1.345, (1.896–6.704), p = 0.001. the experimental group: − 4.52 (− 5.72, − 2.82) (mean change Median
In conclusion, we found moderate evidence in favour of whole-body (IQR)) in the experimental group versus − 0.35 (− 1.18, 1.75) in the
vibration being included in a training program versus the training control group, p < 0.001. Both groups significantly improved gait speed,

143
M. Geerars et al. Gait & Posture 91 (2022) 137–148

Table 3
Risk of bias table and methodologic quality scores of the studies included.

although no significant difference between the groups was found. individually adjusted for each participant to achieve a normal walking
According to the BES, moderate evidence was found for the effec­ pattern. The short-term effects on knee hyperextension in midstance and
tiveness, on degree of knee hyperextension, of prowling and proprio­ the electromyographic activation in the quadriceps and hamstrings were
ceptive training adjusted to routine physiotherapy as compared to measured in one session which entailed 5 min of CTG and 5 min of GTG.
routine physiotherapy in the short-term. We found no evidence for the In the post-stroke group, a significant difference in favour of GTG was
extra benefits of prowling and proprioceptive training on gait speed. found on knee hyperextension in mid-stance (degrees) pre- and
post-testing: 189.1(3.9) (mean(SD)) and 178.1(7.7) p < 0.01.
3.4.1.3. Biofeedback training adjusted to a Motor Relearning Programme We found moderate evidence for effectiveness in favour of GTG as
[35] (MRP) versus MRP. Morris et al. [29] (RCT, n = 26, excellent compared to CTG on knee hyperextension in the short-term (as applied
quality) included participants 1–3 months post-stroke. The effects of in one session) for the post-stroke group.
biofeedback training as an add-on treatment to MRP were investigated.
Participants were included if they were able to safely walk 10 m without 3.4.1.5. Biofeedback training adjusted to conventional physiotherapy
aids or orthotics. Both the experimental and control groups received a versus conventional physiotherapy. Ceceli et al. [32] (CCT, n = 41, fair
MRP [35] for 8 weeks. In this program, gait training was given in which quality) included persons 0 to >1 year post-stroke. Participants had a
stereotyped and abnormal synergistic activity was discouraged. As an normal passive range of motion in the affected lower extremity and were
add-on treatment to the MRP, the experimental group received able to walk unassisted for 50 steps. The experimental group received
biofeedback training: standing and gait training with auditory feedback gait training with the instruction not to hyperextend the knee. Auditory
from an electro-goniometer during the first 4 weeks of therapy. For the feedback on knee hyperextension was provided by using an electrome­
biofeedback training an electro-goniometric device mounted on the chanical rotational goniometer for ten days, 30 min per day, which was
knee of the participant, provided feedback on the joint position of the added to the conventional physiotherapy. The control group only
knee. An acoustic signal was given when the knee extension exceeded a received conventional physiotherapy i.e. exercises for the pelvis, hip
certain angle. Short-term training effects (after 4 and 8 weeks) were control and weight shifting. Researchers measured the number of steps
measured on peak knee hyperextension, gait speed and gait symmetry. with knee hyperextension within 50 steps and gait speed after 10 days
After 8 weeks of training, a significant benefit of biofeedback training and 6 months of training. Between groups, results showed a significant
was found to reduce the peak knee hyperextension (Mann-Whitney test, decrease in the number of steps with knee hyperextension in favour of
p = 0.011). No significant differences between groups were found on the experimental group in the short-term (10 days): the experimental
peak knee hyperextension after 4 weeks of training or on gait speed and group went from 45.19 to 7.538 steps (mean (SD: not reported)) versus
gait symmetry after 4 and 8 weeks of training. the control group that went from 48.67 to 39.38 steps (p < 0.05). No
According to the BES, moderate evidence was found in favour of short-term differences between the groups were found on gait speed.
biofeedback treatment adjusted to a MRP versus MRP alone [35] on The long-term effects (after 6 months) were not statistically analysed
knee hyperextension in the short-term after 8 weeks, although no evi­ because the number of participants that showed up at the check-up was
dence was found after 4 weeks. We found no evidence for an extra too low.
benefit of biofeedback treatment on gait speed. Therefore, we found limited evidence in favour of the biofeedback
treatment adjusted to conventional therapy versus conventional therapy
3.4.1.4. Guidance Tubing Gait on the treadmill versus a conventional on knee hyperextension in the short-term. No evidence was found for the
treadmill gait. Lee et al. [31] (cross-over RCT, n = 18 stroke survivors effectiveness of the interventions on gait speed and gait symmetry in the
and 15 healthy controls, good quality) compared a guided tubing gait short-term.
(GTG) on a treadmill with a conventional treadmill gait (CTG) in stroke
survivors 4–10 months post-stroke, having a Berg Balance Score > 40, 3.4.2. Orthotic treatment
with healthy controls. For the purpose of this review, we only used the Three trials [6,30,33] investigated the benefits of the use of an
data in which both treatments were compared in stroke survivors. An orthosis compared to no orthosis to prevent knee hyperextension. Bae
assistive guidance force using tubing was given to improve knee joint et al. [33] compared FES and orthotic treatments (see 3. FES)
stabilization during mid-stance and restore knee joint muscle imbal­
ances and kinematics in gait. Tubing resistance was applied on the 3.4.2.1. Hinged soft knee orthosis versus no orthosis. Portnoy et al. [30]
popliteal fossa and above the medial malleolus. Previous to the mea­ (cross-over RCT, n = 31, fair quality) investigated the added value of a
surements, each participant underwent 2–5 sessions of 10 min of CTG hinged soft knee orthosis in persons, 3 months to 25 years post-stroke.
and 10 min of GTG to modify the tubing tension: the applied force was All participants walked independently at the time of recruitment and

144
M. Geerars et al. Gait & Posture 91 (2022) 137–148

Table 4 had a paresis or spasticity in the lower limb muscles. An 8-week training
Evidence for effectiveness of treatment on knee hyperextension post-stroke. program was conducted in which the participants wore the orthosis
Treatment Phase post-stroke Evidence for Grade of Gait Gait during the day for 4 weeks; the other 4 weeks, they did not wear the
knee hyperextension in speed symmetry orthosis. Measurements on knee hyperextension, gait speed and gait
stance phase symmetry were conducted without orthosis at baseline, after 4 weeks of
Proprioceptive training wearing the orthosis, and after 4 weeks of not wearing the orthosis.
Training program (a.o.in semi When wearing the orthosis for 4 weeks, knee hyperextension was
squatting position) plus completely prevented: hyperextension (degrees) with versus without
Whole-Body Vibration* vs
training program
orthosis: 10.1 (9.5) (mean(SD)) versus − 8.2 (7.2) p < 0.001. On gait
Subacute phase speed, the authors reported no significant improvement in velocity
Short-term effects (8 weeks) ++ ++ (measured in step length × (number of steps/min)) and a statistically
Prowling and proprioceptive significant improvement on the 10 m-walk-test (10MWT) in favour of
training (both with knee
the orthosis group (21.3(16.6) (mean(SD)) versus 22.9(16.6) for the
flexion) adjusted to routine
physiotherapy* vs routine group without orthosis, p = 0.011). We contacted the authors to get a
physiotherapy better understanding of the results concerning gait speed. We decided to
Acute/subacute phase ++ NE only include the results of the 10MWT because the authors pointed out
Short-term effects (6 sessions) that gait velocity was not measured validly due to the small laboratory.
Biofeedback training (during
walking with attention to a
No significant changes in gait symmetry were found in the short-term.
flexed knee) plus Motor So, according to the BES, in the short-term, limited evidence was
Relearning Program (MRP)* found for the effectiveness of wearing a soft-hinged orthosis on knee
vs MRP hyperextension and gait speed compared to not wearing the orthosis.
Subacute phase ++ NE NE
Short-term effects (8 weeks)
Guidance Tubing Gait 3.4.2.2. Knee Ankle Foot Orthosis (KAFO) versus no orthosis. Boudarham
(assistive guidance force to et al. [6] (CCT, n = 11, fair quality) investigated the use of a KAFO in
knee flexion using tubing) on people 6–56 years post-stroke. All participants were able to walk 10 m
treadmill* vs conventional
without walking aids and exhibited spasticity or weakness in the
treadmill gait
Subacute-chronic phase ++ quadriceps. Each participant performed two gait cycles: without KAFO
Short-term effects (1 session, (control condition), and - after a 10-min rest - with their own KAFO. In
during tubing gait) one session, the degree of knee hyperextension, gait speed and gait
Biofeedback training with symmetry were measured. In the stance phase, the knee hyperextension
attention to a flexed knee*
(degrees) was significantly reduced with a KAFO as compared to no
plus conventional
physiotherapy vs KAFO, although no effect size was reported: with KAFO versus without
conventional physiotherapy. KAFO: − 7.6 (7.4), (mean(SD)) versus − 16.2(11.9), p = 0.029. The knee
Acute-chronic phase hyperextension was not totally resolved: the mean decrease of hyper­
Short-term effects (10 days) NE NE
+
extension with KAFO was around 8◦ . Gait speed (m/sec) increased
Orthotic treatment
Hinged soft knee orthosis* vs significantly in the KAFO condition: with a KAFO versus without a KAFO
no orthosis 0.73(0.34) (mean(SD)) versus 0.57(0.36), p = 0.025. Gait symmetry was
Subacute-chronic phase + + NE measured in both the swing phase and stance phase. In the KAFO con­
Short-term effects (4 weeks) dition, the symmetry between the paretic and the non-paretic limb
Knee ankle foot orthosis
increased significantly during the swing phase: asymmetry ratio with a
(KAFO)* vs no orthosis
Chronic phase KAFO versus without a KAFO 1.27 (0.10) (mean(SD)) versus 1.93 (0.77),
Short-term effects (1 session) + + + (in swing p = 0.014. No significant difference between the two conditions was
phase) found in the stance phase.
NE (in
In conclusion, we found limited evidence for short-term effectiveness
stance
phase)
in favour of a KAFO on knee hyperextension, gait speed and gait sym­
AFO (5◦ ankle dorsiflexion)* vs metry in the swing phase, although not in the stance phase, as measured
no orthosis (barefoot immediately after putting the KAFO on.
walking)
Chronic phase
3.4.3. Functional Electrical Stimulation (FES)
Short-term effects (1 session) NE NE
Functional Electrical In FES, the electrical stimulation of muscles with poor nerve control
Stimulation (FES) is conducted to evoke a contraction and obtain functional movement.
Dorsiflexor FES vs barefoot
walking
3.4.3.1. Dorsiflexor FES versus 5◦ ankle-foot orthosis (AFO) and barefoot
Chronic phase
Short-term effects (in one NE NE walking. Bae et al. [33] (CCT, n = 12, good quality) compared three
session) walking conditions, consecutively barefoot walking, ankle-foot orthosis
Dorsiflexor FES vs AFO (5◦ (AFO), and dorsiflexor FES in a cross-over trial in participants who were
ankle dorsiflexion)
at least six months post-stroke. Comparing the three walking conditions
Chronic phase
Short-term effects (in one NE NE
in all participants, neither significant differences on the knee angle in
session) midstance nor gait speed were found.
So, according to the BES, in the short-term (one session), we found no
*, in favor of.
evidence for the effectiveness of dorsiflexor FES or 5◦ ankle dorsiflexor
+++, strong evidence found; ++, moderate evidence found; +, limited evidence
found. AFO on knee hyperextension in the stance phase or the gait speed as
+/− , conflicting evidence found; NE, no evidence found for the effect of treat­ compared to barefoot walking.
ment; KAFO: knee ankle foot orthosis: AFO; Ankle foot orthosis.

145
M. Geerars et al. Gait & Posture 91 (2022) 137–148

4. Discussion So, interventions on knee hyperextension that are beneficial to energy

cost in gait might improve walking participation and the performance of
This systematic review studied the evidence for the effectiveness of daily activities.
interventions on knee hyperextension in post-stroke gait. Three types of Bleyenheuft et al. [7] concluded that none of the selected articles
interventions were identified: proprioceptive training, orthotic treat­ they included in their literature review (2010) considered the various
ment, and FES. The eight included studies investigated the effectiveness causes of knee hyperextension. They described a proposal for a specific
of an intervention on knee hyperextension in the stance phase of gait. treatment strategy for each cause of knee hyperextension based on their
Seven studies also investigated the influence of a treatment on gait speed clinical experience, in order to help identify the most appropriate
and three investigated gait symmetry. Only short-term effects were re­ treatment. From our point of view, a clinician should start with pro­
ported (from one day to eight weeks). prioceptive training, keeping in mind that in some cases, additional
The adjustment of a form of proprioceptive training to physiotherapy treatment might also be necessary to create better motoric conditions for
training programs with the aim of keeping the knee flexed during ex­ preventing knee hyperextension. These additional interventions could
ercises and walking seems to be effective (moderate evidence) on knee include reducing spasticity or stimulating muscles using FES or even
hyperextension in gait, in the short-term, as applied during the subacute compensating for the deficits such as in orthotic treatment.
phase post-stroke. Neither evidence for the effects on gait speed nor gait It is striking that, until now, in scientific literature no consensus can
symmetry were found for proprioceptive training, although the adjust­ be found as to the different causes of knee hyperextension in people
ment of Whole-Body Vibration had some effect on gait speed (moderate recovering from a stroke described. In 2012, researchers came to an
evidence). Nevertheless, we would expect that the other studies on agreement about the role of the posterior muscles of the lower leg (mm.
proprioceptive training included in this review would have found triceps surae) [7,14,44,45] and weak hip extensor muscles [7,45] in
similar results on gait speed if gait speed was related to the results on causing knee hyperextension. The influence of over-activity or weakness
knee hyperextension. of the knee extensors (m. rectus femoris) as stated by Bleyenheuft [7]
In the three studies we found on orthotic treatment, three different and Fish [45] was contradicted by the findings of Gross [12] and Cooper
orthoses were studied. The highest level of evidence that could be [44]. Gross [12] found that a reduction in the overactivity of the rectus
reached was ‘limited‘. It seems that only orthoses that cover the knee had femoris by a rectus femoris nerve blocker had no influence on knee
some effect on knee hyperextension and gait speed. On symmetry, no hyperextension. Bleyenheuft [7] described the role of weak knee flexors
evidence was found, except for a KAFO in the swing phase. (mm. hamstrings) failing to control knee flexion in the stance phase,
No evidence was found for FES, although this conclusion was based causing knee hyperextension. This was not supported by the findings of
only on one single study. Cooper [46]. Cooper [44] evaluated the association between muscle
The benefits of proprioceptive training on knee hyperextension, as weakness and knee hyperextension. Weak hamstrings and quadriceps
found in this review, were confirmed by the findings of two trials [36, were found to play no role in knee hyperextension in gait. However, in
37] that we had to exclude: Basaglia et al. [36] included both people the study of Lee [31] included in our review, researchers saw that the
with stroke and intracranial pressure and found long-term benefits (one tested intervention had a significant effect of the quadriceps/hamstring
year) on knee hyperextension using biofeedback. McCain et al. [37] muscle imbalance ratio. This may suggest that addressing the coordi­
found better gait kinematics, gait speed and symmetry as well as an nation between quadriceps and hamstring muscles could play a role in
absence of knee hyperextension after gait training with a partial weight the normalization of the neuromuscular imbalance of gait.
bearing system as applied before walking over ground was trained, in the Whether all cases of knee hyperextension should be treated in people
sub-acute phase post-stroke. recovering from a stroke remains a point of discussion. For functional
Within this review, the study of Bae et al. [33] found no evidence for recovery of the ability to walk, recovery of joint kinematics to support a
the benefit of a 5◦ dorsiflexor AFO on knee hyperextension. However, normal walking pattern is not necessary [47]. Several authors described
from a clinical point of view, we know that an AFO is an orthosis that is knee hyperextension as a useful adjustment developed to achieve limb
easy to put on with one hand. Positives signs for findings on other types stability while walking [13,45,47]. Different points of view may be
of AFO in clinical practice can also be found in two other small studies: caused by lack of mid- and long-term evidence for the occurrence of
Kobayashi et al. [21] (n = 6) and Sutdet [38] (n = 4). These studies were secondary knee pain. In general, in an active individual, the possible
excluded from our systematic review because of the small sample-size consequence of knee hyperextension after a stroke may be instability of
(n<10). Kobayashi et al. [21] (n = 6) found a significant reduction of the knee, which is caused by stretching the posterior ligaments and
knee hyperextension using an articulated AFO with adjustable plantar capsule, degenerative joint disease and chronic knee pain [10]. Younger
flexion resistance when the plantar flexion resistance increased. Further, or less severely impaired people with a more favourable walking ability
Sutdet et al. [38] (n = 4) found a significant reduction of knee hyper­ might be more at risk and should definitely be treated, as well as people
extension using a Rigid Tuned AFO inclining the tibia angle from 0◦ to who are overweight and those who must maintain a standing position
15◦ . Therefore, in our opinion, it could be worthwhile to investigate the for a long time, due to the risk of overloading the knee joint [32,36].
benefits of AFO types on a larger scale. However, treatment of knee hyperextension in individuals with a serious
On FES, we could only include the study of Bae et al. [33]. Another neurological outcome, poor walking autonomy, a reduced stance phase
study on FES from Springer et al. [39] also found no benefits of dorsi­ or a low physical activity level, preventing knee hyperextension to avoid
flexor FES on knee hyperextension. However, the latter study was knee pain might be less necessary, because there is a lower risk of
excluded because only seven people with hemiparesis (of which six post overloading the knee joint in the case of incidental use. Tani et al. [13]
stroke) were included. Nevertheless, and maybe interesting to study in suggest secondary knee pain might not be more common in people
future research, Springer et al. [39] found that dual channel FES, acti­ recovering from a stroke who suffer from knee hyperextension as
vating both the peroneal and hamstring muscles decreased knee hy­ compared to the general population. This supports the assumption that
perextension in this group. the need for treatment of the knee hyperextension depends on the in­
Although many researchers [7,15,28,29,32] from the trials included dividual stroke survivor.
in the current review described the problem ‘increased energy cost while The results of this review give us some indications that, initially, the
walking with knee hyperextension’, none of them investigated the effect clinician should attempt to restore knee control of the hyperextended
of the intervention on energy costs. Nevertheless, from a patient’s knee by implementing a training program with a proprioceptive
perspective, energy cost in gait is an important factor in life after a stroke approach. Whole body vibration is an adjustment that seems beneficial
[40]. Higher energy costs of walking is a functional burden and might to gait speed as well. It is unclear if the benefits are maintained in the
limit walking and daily activities related to walking post-stroke [41–43]. long term. Proprioceptive training can be performed in both the (sub)

146
M. Geerars et al. Gait & Posture 91 (2022) 137–148

acute and chronic phases post stroke. In our opinion, in the cases where & editing. Bionka M.A Huisstede: Conceptualization, Methodology,
this retraining method is not successful, treatments might compensate Validation, Formal analysis, Writing - review & editing.
for knee control by using and orthotic treatment, although it is not clear
which orthosis is best. It should be taken into account that wearing an Declaration of Competing Interest
orthosis is not well-accepted by patients and often refused, for cosmetic
reasons or due to discomfort [6,48]. In addition, putting on an orthosis is The authors report no declarations of interest.
often difficult for a person recovering from a stroke due to a paretic arm.
It should be taken into account that the adaptation of gait to the mod­ Acknowledgements
ifications (i.e. decreased hyperextension, increased range of motion of
the ankle) is a long process [49] and thus, a period of training is The authors want to express their gratitude to B.M.R Kramer,
required. FES in the case of muscle weakness or treatment aiming to librarian of the Utrecht University Library, for her advice on the search
reduce spasticity might be beneficial in individual cases, although we string.
found no evidence for it in the present review.
To draw firm conclusions, more studies with larger sample sizes are Appendix A. Supplementary data
needed. Future studies should also report mid- and long-term results in
order to determine if the effects of the intervention are sustainable. The Supplementary material related to this article can be found, in the
influence of these interventions on energy costs should be investigated. online version, at doi:https://doi.org/10.1016/j.gaitpost.2021.08.016.
(Retrospective) research of mid-, and long-term effects of not treating
knee hyperextension or non-adherence to orthosis would underline or References
deny the need for treating knee hyperextension in people experiencing
secondary knee pain. [1] C.O. Johnson, M. Nguyen, G.A. Roth, E. Nichols, T. Alam, D. Abate, et al., Global,
regional, and national burden of stroke, 1990–2016: a systematic analysis for the
Global Burden of Disease Study 2016, Lancet Neurol. 18 (2019) 439–458, https://
4.1. Study limitations doi.org/10.1016/S1474-4422(19)30034-1.
[2] G.J. Hankey, Stroke, Lancet 389 (2017) 641–654, https://doi.org/10.1016/S0140-
The current review is the first systematic literature review on the 6736(16)30962-X.
[3] C.J. Winstein, J. Stein, R. Arena, B. Bates, L.R. Cherney, S.C. Cramer, F. Deruyter, J.
effectiveness of interventions on knee hyperextension in post-stroke
J. Eng, B. Fisher, R.L. Harvey, C.E. Lang, M. MacKay-Lyons, K.J. Ottenbacher,
gait. The review was conducted following the steps of the PRISMA-P S. Pugh, M.J. Reeves, L.G. Richards, W. Stiers, R.D. Zorowitz, Guidelines for adult
2015 statement [17]. However, some limitations should be addressed. stroke rehabilitation and recovery, Stroke 47 (2016), https://doi.org/10.1161/
STR.0000000000000098.
Only five out of eight selected trials had a randomized study-design.
[4] A. Soundy, C. Liles, B. Stubbs, C. Roskell, Identifying a framework for hope in order
Therefore, the results needed to be treated with caution because of the to establish the importance of generalised hopes for individuals who have suffered
risk of bias in the analysis. In most studies, the sample sizes were low. a stroke, Adv. Med. 2014 (2014), 471874, https://doi.org/10.1155/2014/471874.
Selection bias may have occurred due to omitting two full-text publi­ [5] B. Balaban, F. Tok, Gait disturbances in patients with stroke, PM&R 6 (2014)
635–642, https://doi.org/10.1016/j.pmrj.2013.12.017.
cations [26,27] that we had to exclude because the full text was not [6] J. Boudarham, R. Zory, F. Genet, G. Vigné, D. Bensmail, N. Roche, D. Pradon,
available [26], or was only available in Chinese [27]. A modified Downs Effects of a knee-ankle-foot orthosis on gait biomechanical characteristics of
and Black checklist [18] was used to assess the methodological quality. paretic and non-paretic limbs in hemiplegic patients with genu recurvatum, Clin.
Biomech. (Bristol, Avon). 28 (2013) 73–78, https://doi.org/10.1016/j.
The ratings “excellent”, “good”, “fair” or “poor” are arbitrary, but clinbiomech.2012.09.007.
nevertheless, frequently used in similar reviews [19,20,22,23]. A con­ [7] C. Bleyenheuft, Y. Bleyenheuft, P. Hanson, T. Deltombe, Treatment of genu
flict of interest must be considered in the trial of Portnoy [30] as it was recurvatum in hemiparetic adult patients: a systematic literature review, Ann.
Phys. Rehabil. Med. 53 (2010) 189–199, https://doi.org/10.1016/j.
funded by an orthotic company. rehab.2010.01.001.
[8] M. Appasamy, M.E. De Witt, N. Patel, N. Yeh, O. Bloom, A. Oreste, Treatment
5. Conclusion strategies for genu recurvatum in adult patients with hemiparesis: a case series,
PM&R 7 (2015) 105–112, https://doi.org/10.1016/j.pmrj.2014.10.015.
[9] S.J.’ Olney, T.N. Monga, P.A. Costigan, Mechanical Energy of Walking of Stroke
Studies reporting on the effectiveness of proprioceptive training, Patients, (n.d.). http://www.archives-pmr.org/article/0003-9993(86)9010
orthotic treatment, and FES to treat knee hyperextension in post-stroke 9-7/pdf.
[10] J.K. Loudon, H.L. Goist, K.L. Loudon, Genu recurvatum syndrome, J. Orthop. Sport.
gait were found. Promising results (moderate evidence) were found for
Phys. Ther. 27 (1998) 361–367, https://doi.org/10.2519/jospt.1998.27.5.361.
some “proprioceptive approaches” as an add-on therapy to physio­ [11] D.J, J. Perry, Gait analysis: normal and pathological function, J. Pediatr. Orthop.
therapy training programs on knee hyperextension in the subacute (November (1)) (1992) 815.
phase post-stroke in the short-term. The effects of proprioceptive [12] R. Gross, L. Delporte, L. Arsenault, P. Revol, M. Lefevre, D. Clevenot, D. Boisson,
P. Mertens, Y. Rossetti, J. Luauté, Does the rectus femoris nerve block improve
training on gait speed need further investigation; no evidence for ben­ knee recurvatum in adult stroke patients? A kinematic and electromyographic
efits on gait-symmetry was found. For the effectiveness of orthotics and study, Gait Posture 39 (2014) 761–766, https://doi.org/10.1016/j.
FES, only limited and no evidence in the short-term was found. None of gaitpost.2013.10.008.
[13] Y. Tani, Y. Otaka, M. Kudo, T. Kurayama, K. Kondo, Prevalence of genu recurvatum
the included studies reported on mid- and long-term effects. More high- during walking and associated knee pain in chronic hemiplegic stroke patients: a
quality studies concentrating on treating knee hyperextension in post- preliminary survey, J. Stroke Cerebrovasc. Dis. 25 (2016) 1153–1157, https://doi.
stroke gait, that not only report short-term, but also mid- and long- org/10.1016/j.jstrokecerebrovasdis.2016.01.028.
[14] J.S. Higginson, F.E. Zajac, R.R. Neptune, S.A. Kautz, C.G. Burgar, S.L. Delp, Effect
term effects are definitely needed. of equinus foot placement and intrinsic muscle response on knee extension during
stance, Gait Posture 23 (2006) 32–36, https://doi.org/10.1016/j.
Funding gaitpost.2004.11.011.
[15] K.K. Dalal, A.M. Joshua, A. Nayak, P. Mithra, Z. Misri, B. Unnikrishnan,
Effectiveness of prowling with proprioceptive training on knee hyperextension
This research did not receive any specific grant from funding among stroke subjects using videographic observation- a randomised controlled
agencies in the public, commercial or non-for-profit sectors. trial, Gait Posture 61 (2018) 232–237, https://doi.org/10.1016/j.
gaitpost.2018.01.018.
[16] A.C. Novak, S.J. Olney, S. Bagg, B. Brouwer, Gait changes following botulinum
CRediT authorship contribution statement toxin a treatment in stroke, Top. Stroke Rehabil. 16 (2009) 367–376, https://doi.
org/10.1310/tsr1605-367.
Marieke Geerars: Conceptualization, Methodology, Validation, [17] D. Moher, L. Shamseer, M. Clarke, D. Ghersi, A. Liberati, M. Petticrew, P. Shekelle,
L.A. Stewart, P.-P. Group, Preferred reporting items for systematic review and
Formal analysis, Investigation, Writing - original draft. Nympha meta-analysis protocols (PRISMA-P) 2015 statement, Syst. Rev. 4 (2015) 1, https://
Minnaar-van der Feen: Formal analysis, Investigation, Writing - review doi.org/10.1186/2046-4053-4-1.

147
M. Geerars et al. Gait & Posture 91 (2022) 137–148

[18] S.H. Downs, N. Black, The feasibility of creating a checklist for the assessment of [34] J.E. Aman, N. Elangovan, I.L. Yeh, J. Konczak, The effectiveness of proprioceptive
the methodological quality both of randomised and non-randomised studies of training for improving motor function: a systematic review, Front. Hum. Neurosci.
health care interventions, J. Epidemiol. Commun. Health 52 (1998) 377–384, 8 (2015), https://doi.org/10.3389/fnhum.2014.01075.
https://doi.org/10.1136/JECH.52.6.377. [35] J. Carr, R. Shepherd, A Motor Relearning Programme for Stroke, 1982. https://sch
[19] S.R. O’Connor, M.A. Tully, B. Ryan, J.M. Bradley, G.D. Baxter, S.M. McDonough, olar.google.nl/scholar?hl=nl&as_sdt=0%2C5&q=a+motor+relearning+progr
Failure of a numerical quality assessment scale to identify potential risk of bias in a amme+for+stroke+carr&btnG=.
systematic review: a comparison study, BMC Res. Notes 8 (2015) 224, https://doi. [36] N. Basaglia, N. Mazzini, P. Boldrini, P. Bacciglieri, E. Contenti, G. Ferraresi,
org/10.1186/s13104-015-1181-1. Biofeedback treatment of genu-recurvatum using an electro goniometric device
[20] J. Galvin, R. McDonald, C. Catroppa, V. Anderson, Does intervention using virtual with an acoustic signal. One-year follow-up, Scand. J. Rehabil. Med 21 (1989)
reality improve upper limb function in children with neurological impairment: a 125–130. https://www.embase.com/search/results?subaction=viewreco
systematic review of the evidence, Brain Inj. 25 (2011) 435–442, https://doi.org/ rd&id=L19237617&from=export.
10.3109/02699052.2011.558047. [37] K.J. McCain, F.E. Pollo, B.S. Baum, S.C. Coleman, S. Baker, P.S. Smith, Locomotor
[21] T. Kobayashi, M.S. Orendurff, M.L. Singer, F. Gao, W.K. Daly, K.B. Foreman, treadmill training with partial body-weight support before overground gait in
Reduction of genu recurvatum through adjustment of plantarflexion resistance of adults with acute stroke: a pilot study, Arch. Phys. Med. Rehabil. 89 (2008)
an articulated ankle-foot orthosis in individuals post-stroke, Clin. Biomech. 35 684–691, https://doi.org/10.1016/j.apmr.2007.09.050.
(2016) 81–85, https://doi.org/10.1016/j.clinbiomech.2016.04.011. [38] P. Sutdet, N. Polhan, V. Prakotmongkol, J. Poomulna, F. Trifani, T. Iswana, L.
[22] J. Eng, R. Teasell, W. Miller, D. Wolfe, A. Townson, J.-A. Aubut, C. Abramson, K. Nipuni, M. De Silva, S. Tharin, The effect of tibia inclination on gait pattern of
J. Hsieh, S. Connolly, K. Konnyu, Spinal cord injury rehabilitation evidence: stroke patients with genu recurvatum who wear an Ankle Foot Orthosis: a pilot
method of the SCIRE systematic review, Top. Spinal Cord Inj. Rehabil. 13 (2007) study, Technol. Disabil. 31 (2019) 77–82, https://doi.org/10.3233/TAD-180211.
1–10, https://doi.org/10.1310/sci1301-1. [39] S. Springer, J.-J. Vatine, A. Wolf, Y. Laufer, The effects of dual-channel functional
[23] A.R. Marinho-Buzelli, A.M. Bonnyman, M.C. Verrier, The effects of aquatic therapy electrical stimulation on stance phase sagittal kinematics in patients with
on mobility of individuals with neurological diseases: a systematic review, Clin. hemiparesis, J. Electromyogr. Kinesiol. 23 (2) (2013) 476–482, https://doi.org/
Rehabil. 29 (2015) 741–751, https://doi.org/10.1177/0269215514556297. 10.1016/j.jelekin.2012.10.017.
[24] M. van Tulder, A. Furlan, C. Bombardier, L. Bouter, Updated method guidelines for [40] A. Pollock, B. St George, M. Fenton, L. Firkins, Top ten research priorities relating
systematic reviews in the Cochrane collaboration back review group, Spine (Phila. to life after stroke, Lancet Neurol. 11 (2012) 209, https://doi.org/10.1016/S1474-
Pa. 1976) 28 (2003) 1290–1299, https://doi.org/10.1097/01. 4422(12)70029-7.
BRS.0000065484.95996.AF. [41] A. Danielsson, K.S. Sunnerhagen, Energy expenditure in stroke subjects walking
[25] S.P. Hombergen, B.M. Huisstede, M.F. Streur, H.J. Stam, J. Slaman, J.B. Bussmann, with a carbon composite ankle foot orthosis, J. Rehabil. Med. 36 (2004) 165–168,
R.J. van den Berg-Emons, Impact of cerebral palsy on health-related physical https://doi.org/10.1080/16501970410025126.
fitness in adults: systematic review, Arch. Phys. Med. Rehabil. 93 (2012) 871–881, [42] J.A.M. Ribeiro, S.G. Oliveira, L. Di Thommazo-Luporini, C.I. Monteiro, S.
https://doi.org/10.1016/j.apmr.2011.11.032. A. Phillips, A.M. Catai, A. Borghi-Silva, T.L. Russo, Energy cost during the 6-minute
[26] R.E. Hogue, S. McCandless, Genu recurvatum: auditory biofeedback treatment for walk test and its relationship to real-world walking after stroke: a correlational,
adult patients with stroke or head injuries, Arch. Phys. Med. Rehabil. 64 (1983) cross-sectional pilot study, Phys. Ther. 99 (12) (2019) 1656–1666, https://doi.org/
368–370. http://www.ncbi.nlm.nih.gov/pubmed/6882176. 10.1093/ptj/pzz122.
[27] S.-H. Gao, C.-M. Ni, R. Han, C. Li, X.-H. Wu, Preventive and therapeutic effect of [43] J.C. Polese, L. Ada, L.F. Teixeira-Salmela, Relationship between oxygen cost of
early isolated and resisted movement training on stroke genu recurvatum and walking and level of walking disability after stroke: an experimental study,
hemiplegic gait, Chin. J. Clin. Rehabil. 10 (September (25)) (2006) 33–35. Physiother. Res. Int. 23 (2018), e1688, https://doi.org/10.1002/pri.1688.
[28] C. Guo, X. Mi, S. Liu, W. Yi, C. Gong, L. Zhu, S. Machado, T.-F. Yuan, C. Shan, [44] A. Cooper, G.A. Alghamdi, M.A. Alghamdi, A. Altowaijri, S. Richardson, The
Whole body vibration training improves walking performance of stroke patients relationship of lower limb muscle strength and knee joint hyperextension during
with knee hyperextension: a randomized controlled pilot study, CNS Neurol. the stance phase of gait in hemiparetic stroke patients, Physiother. Res. Int. 17
Disord. Drug Targets 14 (2015) 1110–1115. http://www.ncbi.nlm.nih.gov/pub (2012) 150–156, https://doi.org/10.1002/pri.528.
med/26556074. [45] D.J. Fish, Genu recurvatum: identification of three distinct mechanical profiles,
[29] M.E. Morris, T.A. Matyas, T.M. Bach, P.A. Goldie, Electrogoniometric feedback: its J. Prosthet. Orthot. 10 (1998) 26.
effect on genu recurvatum in stroke, Arch. Phys. Med. Rehabil. 73 (1992) [46] A. Cooper, S. Richardson, G.A. Alghamdi, A. Williams, C. Butterworth, A. Griffiths,
1147–1154. http://www.ncbi.nlm.nih.gov/pubmed/1463378. Assessing knee hyperextension in stroke patients ‘expert clinician compared to
[30] S. Portnoy, A. Frechtel, E. Raveh, I. Schwartz, Prevention of genu recurvatum in siliconcoach’, Int. J. Stroke 5 (2010) 58.
poststroke patients using a hinged soft knee orthosis, PM&R 7 (2015), https://doi. [47] R.B. Huitema, A.L. Hof, T. Mulder, W.H. Brouwer, R. Dekker, K. Postema,
org/10.1016/j.pmrj.2015.04.007. Functional recovery of gait and joint kinematics after right hemispheric stroke,
[31] J.J. Lee, J. (Sung) H. You, Effects of novel guidance tubing gait on Arch. Phys. Med. Rehabil. 85 (2004) 1982–1988. http://www.ncbi.nlm.nih.gov/p
electromyographic neuromuscular imbalance and joint angular kinematics during ubmed/15605337.
locomotion in hemiparetic stroke patients, Arch. Phys. Med. Rehabil. 98 (2017) [48] J. Leung, A. Moseley, Impact of ankle-foot orthoses on gait and leg muscle activity
2526–2532, https://doi.org/10.1016/j.apmr.2017.04.018. in adults with hemiplegia, Physiotherapy 89 (2003) 39–55, https://doi.org/
[32] E. Ceceli, E. Dursun, A. Cakci, Comparison of joint-position biofeedback and 10.1016/S0031-9406(05)60668-2.
conventional therapy methods in genu recurvatum after stroke - 6 months’ follow- [49] L. Terreaux, R. Gross, F. Leboeuf, H. Desal, O. Hamel, J.P. Nguyen, et al., Benefits
up, Eur. J. Phys. Med. Rehabil 6 (1996) 141–144. https://www.embase.com/sea of repetitive transcranial magnetic stimulation (rTMS) for spastic subjects: clinical,
rch/results?subaction=viewrecord&id=L26394465&from=export. functional, and biomechanical parameters for lower limb and walking in five
[33] D.-Y. Bae, J.-H. Shin, J.-S. Kim, Effects of dorsiflexor functional electrical hemiparetic patients, Sci. World J. 2014 (2014), 389350, https://doi.org/10.1155/
stimulation compared to an ankle/foot orthosis on stroke-related genu recurvatum 2014/389350.
gait, J. Phys. Ther. Sci. 31 (2019) 865–868, https://doi.org/10.1589/jpts.31.865.

148