Clinical Rehabilitation 2010; 24: 810–821

Butler’s neuromobilizations combined with proprioceptive

neuromuscular facilitation are effective in reducing
of upper limb sensory in late-stage stroke subjects:
a three-group randomized trial
Tomasz Wolny, Edward Saulicz, Rafal- Gnat and Mirosl-aw Kokosz Department of Physiotherapy, Chair of Ergonomics,
Prosthetics and Orthotics and Chair of Basics of Physiotherapy, University School of Physical Education, Katowice, Poland

Received 19th February 2009; returned for revisions 22nd February 2009; revised manuscript accepted 14th January 2010.

Question: Are Butler’s neuromobilizations combined with proprioceptive neuromus-

cular facilitation and traditional post-stroke therapy more effective in reducing
affected upper extremity sensory deficits in late-stage stroke subjects than proprio-
ceptive neuromuscular facilitation combined with traditional therapy or traditional
therapy alone?
Design: Pretest–posttest three-group randomized clinical experimental design.
Participants: A total of 96 late-stage stroke subjects were randomly assigned to
three groups.
Intervention: The therapeutic programme in the control group was based on tradi-
tional post-stroke methods. The second group (experimental 1) received in addition
individual therapy based on the proprioceptive neuromuscular facilitation method.
The third group (experimental 2) received a combination: traditional therapeutic
programme plus individual proprioceptive neuromuscular facilitation exercises plus
neuromobilization of the affected upper extremity. All groups received 18 training
sessions lasting about 45 minutes each.
Outcome measures: Assessment of the two-point discriminatory sense (distance
between the tips of the compass when the subject indicated two-point sensation),
stereognosia (identification up to 10 objects by touch) and thermaesthesia (using
hot and cold cylinders on dermatomes C6–C8) were performed.
Results: Analysis of change scores showed that two-point discriminatory sense for
experimental group 2 was significantly better than that in the two other groups
(P50.001). Similar results were registered for thermaesthesia (experimental 2
versus experimental 1 P50.01; experimental 2 versus control P50.001). For
stereognosia the only significant difference was found between experimental
group 2 and the control group (P50.05).
Conclusion: In our subjects, application of Butler’s neuromobilizations combined
with proprioceptive neuromuscular facilitation showed greater effectiveness in
reducing sensory deficits than proprioceptive neuromuscular facilitation or tradi-
tional therapy alone.

Address for correspondence: Miroslaw Kokosz,

ul. Mikolowska 72b, Katowice 40-165, Poland.
e-mail: m.kokosz@awf.katowice.pl
ß The Author(s), 2010.
Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav 10.1177/0269215510367561
 Butler’s method in late-stage stroke subjects 811

Introduction re-education. Using the neuromuscular re-educa-

tion approach but not including the peripheral
The literature indicates that stroke rehabilitation is nerve in the rehabilitation programme seems
based on motor re-education and compensation of inconsistent. Although it is mainly understood
as an entity affecting the central nervous
motor deficits, although hemiplegia selectively
system,7,9,10,13 stroke itself and the subsequent
affecting effectors of the locomotory system is
hemiplegia distort interneuronal and extraneuro-
very rare. Sensory deficits accompanying the
nal mechanics as well. Immobilization of the upper
paralysis are as frequent as post-stroke motor dys-
limb is the main factor in this process. Frequently,
functions.1–4 Stroke interferes with the reception
pathogenic alignments of the limb can develop in
and interpretation of all kinds of peripheral stim-
this way, which may cause misinterpretation of
uli, particularly exteroception and proprioception.
sensory information input on the functional
Proprioceptive impairment represents the greatest
status of the nerve. A similar role may be ascribed
obstacle to proper motor re-education, since sen-
to misleading data coming from the self-innerva-
sory information provided by proprioceptors and
tion system of the nerve, the so-called ‘nervi
transmitted via afferent neural pathways forms the
nervorum’. Disturbed axonal transport also
basis for the cortical motor patterns that are
contributes to defective motor control of the
reflected in muscle behaviour.5 Neuromuscular
extremity.4,14,15
re-education is deeply rooted in proper proprio-
Based on these issues and on ethical consider-
ceptive stimulation and this process is often
ations, the overall research problem was formu-
called the proprioceptive facilitation.6 The situa-
lated as follows: are Butler’s neuromobilizations
tion worsens when other sensory deficits, espe-
combined with the proprioceptive neuromuscular
cially exteroceptive ones, are considered. In the
facilitation method and traditional post-stroke
case of an upper extremity this is particularly
therapy more effective in reducing affected upper
important because the second major function of
extremity sensory deficits in late-stage stroke sub-
the upper limb (apart from motor activity) is to
jects than proprioceptive neuromuscular facilita-
deal with the large number of sensations (i.e.
tion combined only with traditional therapy or
extreroception, thermaesthesia, stereognosia, etc.).
traditional therapy alone? Specific research
Although many approaches to the rehabilitation
questions concerned whether Butler’s neuro-
of motor functions of the affected upper limb are
mobilizations are more effective in the case of
available, there is a significant shortage of thera-
the two-point discriminatory sense, in the case of
peutic tools aiming to improve sensory deficits. It
thermaesthesia and in the case of stereognosia.
seems that exteroceptive disturbances exert some
influence on the functional capacity of the affected
upper extremity, especially the hand.7 Kim and
Choi-Kwon,8 who evaluated discriminatory sense Method
and gnosia in stroke patients, report that the first
entity was diminished in 57 of 67 subjects, and the Design
second entity in 17 of 39 subjects. Although some A pretest–posttest three-group randomized clin-
authors have struggled with the problem of post- ical experimental design was used. Design and
stroke sensory deficits,9–11 their reports provide flow of participants through the trial is outlined
little information on the rehabilitation regimes in Figure 1. The study was approved by the
used. Overall, the literature on treating post- Institutional Biomedical Research Committee
stroke sensory impairments is sparse.12 and all procedures conformed with the
The application of neural tissue mobilizations Declaration of Helsinki of 1983. All subjects
in post-stroke patients may seem controversial. were informed about the aims of the study and
The idea stems from the shortage of therapeutic the experimental procedures, and all gave written
tools aiming to reduce stroke-related sensory informed consent prior to participation. They
deficits, and from general rules of ‘sensory’ reha- were free to withdraw participation at any stage
bilitation based on principles of neuromuscular of the study.
812 T Wolny et al.

Patients referred with stroke history

(n= 164)

Ineligible (n= 68)

Eligible and elected to participate

(n= 96)

Week
Measured two-point discriminatory sense, stereognosia, thermaesthesia
0 Randomized (n= 96)
(n = 32) (n= 32) (n= 32)

Experimental 1 Experimental 2
Control group Traditional therapy plus Traditional therapy plus
Traditional therapy individual therapy based individual therapy based on
on PNF method PNF method plus Butler’s
neuromobilizations
3
Measured two-point discriminatory sense, stereognosia, thermaesthesia
(n= 32) (n= 32) (n = 32)

Figure 1 Design and flow of participants through the trial.

Participants statistical power of ANOVA (probability of com-

A total of 96 stroke subjects participated. The
following selection criteria were applied: post-
stroke period of at least one year duration (late
post-stroke stage), no severe contraindications for
physical loading, ability of independent locomo-
tion, appropriate contact with the subject (possibil-
ity of complete understanding of the objectives and
achieving proper cooperation during therapy). The
first author checked patient history against the
inclusion criteria. Functional status of the patients,
presence of indications and contraindications for
physical loading (patients with contraindications
were excluded) were assessed immediately after
admission to the unit by a physician of internal
medicine, a neurologist and a physiotherapist
(only the physiotherapist – the first author – was
aware of the objective of the study). Subjects meet-
ing the criteria were randomly divided into three
groups: control group, proprioceptive neuromus-
cular facilitation group (experimental 1) and neu-
romobilization (experimental 2). Randomization
was generated by the third author who was blind
to the outcomes of initial assessment.
Based on the results obtained during the prepa-
ratory stage of the experiment we estimated a
sample size necessary to achieve acceptable
mitting the statistical error of the second kind). It
was estimated that groups consisting of about 30
subjects would be sufficient. For allocation pur-
poses we used the method of simple randomiza-
tion. We prepared 96 pieces of paper which
signified individual subjects (in each group we
added 2 subjects to the necessary 30), put them
into container and then drew out one by one.
The order of allocation to groups was: first sub-
ject – proprioceptive neuromuscular facilitation
group (experimental 1), second subject – control
group, third subject – neuromobilization group
(experimental 2), and so on. This order was also
randomized using the same method. No drop-outs
occurred during the entire study.
All diagnostic and therapeutic procedures were
performed between 2000 and 2003 in the District
Hospital in Strzelce Opolskie, Neurological
Rehabilitation Unit, Zawadzkie, Poland.
Intervention
During the 21-day follow-up (regular hospitali-
zation time in the Neurological Rehabilitation
Unit for late-stage stroke patients) all subjects
underwent intensive complex rehabilitation treat-
ment. All groups participated in training protocols

that were standardized in relation to amount of
practice provided by the first author or his gradu-
ate physiotherapy assistants under supervision.
Because of large differences in the functional
capacity of the subjects significant individualiza-
tion of the therapeutic regime was allowed, espe-
cially when the proprioceptive neuromuscular
facilitation method was in use. However, the time-
frame of the single training session always remained
the same. There were 18 sessions performed every
day (except Sundays) between 8 and 12 o’clock.
A single session lasted about 45 minutes.
Control group
The control group received traditional post-
stroke therapy. The therapeutic programme
included individual kinesiotherapy (to reduce
pain, improve stability and control of the joint,
promote functional activity) and additional treat-
ment in the form of: selected physical modalities
(mainly to reduce pain, achieve relaxation and
increased blood flow through tissues), occupa-
tional therapy with self-care elements (to promote
functional daily life activity), orthopaedic supply
(to improve stability and locomotion), logopedia
classes (to reduce speech problems) and psycho-
logical help (to take care of the affected psyche
of the subjects).
Methods of individual kinesiotherapy were
applied as follows:
 passive exercises: 20–30 repetitions for each
joint of affected upper extremity, performed
every day;
 self-assisted exercises (subjects assisted the
motion of their affected upper limbs with
intact upper limbs) and active movements
with no gravity influence (in slings): only for
the affected shoulder joint, 10 minutes, every
day;
 active exercises and active synergistic exercises:
(10 minutes, every day)
 means aiming to improve locomotion and bal-
ance: (15 minutes, every day)
 exercises developing manipulative function of
the hand were also applied (10 minutes, every
day).16
Butler’s method in late-stage stroke subjects 813
Physical therapy modalities included:
 diadynamic current: dosage was adjusted to
individual sensations (clear, not painful), 10
minutes, 10 sessions, every second day17–19;
 interferential current: basal frequency applied
was 50 Hz, spectrum – 50 Hz, 10 minutes, 10
sessions, every day17–19;
 ultrasound: continuous wave 0.5–0.8 W/cm2
applied with mobile probe, 6–8 minutes, 6–8
sessions, every day or every second day17–19;
 cryotherapy: a ‘NR–2’ device (Metrum Cryo
Flex, Blizne Laszczyński, Poland) using
carbon dioxide as the medium, 3 minutes,
15 sessions, every day.17–19
Experimental group 1
Experimental group 1 received individual kine-
siotherapy based on the proprioceptive neuromus-
cular facilitation method.6,20–22 No traditional
individual kinesiotherapy methods were applied
in the control group. In proprioceptive neuromus-
cular facilitation there is no constant therapeutic
algorithm. Specific starting positions, basic princi-
ples, special techniques, applied motion patterns
and overall rehabilitation programme depend on
the patient’s fitness, which is evaluated during
functional assessment. Because of the large differ-
ences in the functional capacity of the subjects sig-
nificant individualization of the therapeutic regime
was allowed. The starting positions used most
likely were: prone, supine and side-lying, four-
point kneeling, kneeling, standing and sitting. In
these positions proprioceptive neuromuscular
facilitation motion patterns for the scapula and
the pelvis (separately and in combination), ipsilat-
eral upper extremity patterns and trunk patterns
incorporating extension were applied. The follow-
ing additional techniques were also frequently
used: combinations of isotonic contraction, stabi-
lizing reversals and replication. Basic principles
such as approximation, compression, irradiation
as well as time and space summation were also
respected. All these techniques, principles and pat-
terns aimed to facilitate specific daily living func-
tions.6,20,21,22 An additional objective was to
reduce pain within the ‘hemiplegic shoulder’.
814 T Wolny et al.
Experimental group 2
Experimental group 2 received individual
kinesiotherapy based on the proprioceptive neuro-
muscular facilitation method (as in experimental
group 1) plus Butler’s neuromobilization directed
to the peripheral nerves of the impaired upper
limb: the median, radial and ulnar nerve.6,14,20
Time of application of proprioceptive neuromus-
cular facilitation exercises was slightly reduced
(about 5–7 minutes) so that together with neuro-
mobilization the total length of the session
remained the same.
Additional treatments
In experimental groups 1 and 2 the additional
treatment applied was as in the control group
(selected physical therapy modalities, occupational
therapy with self-care elements, orthopaedic
supply, logopedia classes, psychological help).
The only inter-group discriminating factor was in
the methods of kinesiotherapy applied.
 Median nerve mobilization: The median nerve
was put under tension by retraction and depres-
sion of the scapula, extension and external rota-
tion of the shoulder joint, extension of the
elbow, supination of the forearm, extension
and ulnar deviation of the carpal joint, exten-
sion of the fingers and rotation/lateral bend of
the cervical spine towards the opposite side.
The mobilization was applied in supine posi-
tion. Tension was increased either by deeper
extension of the fingers and the wrist, or exten-
sion of the elbow, or rotation/lateral bend of
the cervical spine (depending on the location
of the symptoms). In all neuromobilizations
applied, the basic rule was followed to increase
the tension in the most distant-from-symptoms
joint possible. A total of 60 mobilizing
motions were repeated slowly and rhythmically,
with maximal tension lasting about 1 second
and without evoking any pain. The same
methodology was also used for other
neuromobilizations.14,16,23,24
 Radial nerve mobilization: The radial nerve was
put under tension by retraction and depression
of the scapula, extension, abduction and
internal rotation of the shoulder joint, exten-
sion of the elbow, pronation of the forearm,
flexion and ulnar deviation of the wrist, flexion
of the fingers and rotation/lateral bend of the
cervical spine towards the opposite side.
Tension was increased either by deeper flexion
of the fingers and the wrist, or extension of the
elbow, or rotation/lateral bend of the cervical
spine (depending on the location of the
symptoms).14,16,23,24
 Ulnar nerve mobilization: The ulnar nerve was
put under tension by retraction and depression
of the scapula, extension, abduction and inter-
nal rotation of the shoulder joint, flexion of the
elbow, pronation of the forearm, extension and
radial deviation of the wrist, extension of the
fingers and rotation/lateral bend of the cervical
spine towards the opposite side. Tension was
increased either by deeper flexion of the fingers
and the wrist, or extension of the elbow, or rota-
tion/lateral bend of the cervical spine (depend-
ing on the location of the symptoms).14,16,23,24
Outcome measures
The measurements were performed before inter-
vention (initial–week 0) and after its completion
(final–week 3). Because perception and registra-
tion of all investigated sensations are entirely
dependent on the patient, the assessor (first
author) was not blinded to the clinical information
or individual results. In the case of the additional
problem of aphasia, the subjects responded by
nodding or shaking their heads.
Assessment of the two-point discriminatory
sense was performed on the palmar side of the
distal phalangae from I to V. During the initial
measurement the healthy limb was tested prior to
the affected one. No data on the healthy limb were
recorded in the final measurement. The subject
was lying supine with closed eyes. A compass
with two blunted tips and a calliper gauge were
used. The distance between the tips of the compass
(mm) was measured when the subject clearly indi-
cated a two-point sensation. The tips were held
perpendicular to the long axis of a given phalanx
and slight pressure was applied to them. The pres-
sure was strong enough (but still gentle) to feel

stimulation without any pain. The tip-to-tip dis-
tance was gradually increased in steps of 1 mm
starting from 0. This procedure was repeated
several times (with only short breaks to adjust
the compass) and was terminated when a clear
two-point sensation was indicated. Steps from
zero to this point were repeated three times.
Normally, this is enough to reach a steady two-
point sensation level.25–28 From these three
repeated measurements the two values closest to
each other were averaged and analysed. In the sit-
uation where tip-to-tip distance was larger than
the phalanx’s width, the position of the tips was
changed to lie parallel to the long axis of the pha-
lanx. If it was still larger than the distal phalanx’s
length, the medial, or proximal one was tested
instead. For the two-point discriminatory sense
results obtained for individual fingers were aver-
aged their mean value characterized the global
level of the two-point discriminatory sense in
both healthy and affected upper limbs.
In assessing of stereognosia, the subjects were
asked to identify up to 10 commonly used objects
by touch. These were: a comb, a box of matches, a
coin, spectacles, a watch, scissors, a cup, a pen, a
button and a hammer. When independent grasp-
ing was impossible the objects were passively
placed into the subject’s hand. They were hidden
from the subject’s sight. Recognition was con-
firmed by indicating objects on a set of pictures.
A correct response was scored with 10 points, a
false response with 0 points (maximal score ¼ 100
points).29,30 Because of a considerable learning
effect it was not possible to use the same set of
objects for the opposite extremity, therefore no
data on the healthy limb were recorded in the
case of stereognosia.
Thermaesthesia was assessed using two identical
measurement cylinders: one filled with ice, the
other with hot water. During the initial measure-
ment the forearm and fingers of the healthy upper
limb were tested first. No data on the healthy limb
were recorded in the final measurement. Three
dermatomes were assessed: C6 dermatome com-
prising the radial aspect of the forearm, first and
second fingers; C7 dermatome comprising the
dorsal aspect of the forearm, second and third fin-
gers; C8 dermatome comprising the ulnar aspect of
Butler’s method in late-stage stroke subjects 815
the forearm, fourth and fifth fingers.17,23 The sub-
jects assumed a relaxed supine position with closed
eyes. The task was to answer a simple question (i.e.
whether the cylinder touching the skin is hot or
cold). A correct response scored 1 point, a false
response scored 0. The cylinders were applied in
random order three times. Two repeated responses
were recorded as the outcome. Outcomes for the
three investigated dermatomes in the evaluation of
thermaesthesia were summarized, providing a
total outcome expressed in points (range 0–3).
Data analysis
One-way ANOVA and chi-squared test were
used for homogeneity testing. Mixed model of
ANOVA (independent factor: group (control,
experimental 1, experimental 2); repeated factor:
measurement (healthy limb initial, affected limb
initial, affected limb final)) together with a post-
hoc Tukey test were used to evaluate inter- and
intra-group differences in the case of two-point
discriminatory sense and thermaesthesia. Results
obtained for the healthy extremities were regarded
as independent groups.
In the case of the stereognosia evaluation, in
which it was impossible to collect data from the
healthy limbs, the mixed model of ANOVA was
reduced to the form: independent factor: group
(control, experimental 1, experimental 2); repeated
factor: measurement (affected limb initial, affected
limb final).
Change scores were analysed using simple one-
way ANOVA with independent factor: group
(control, experimental 1, experimental 2).
The level of significance (P-value) was set at
0.05. Comparison of initial and final inter-group
differences was used to answer our research ques-
tions. For the two-point discriminatory sense and
thermaesthesia healthy limbs were regarded as
additional reference points.
Results
Characteristics and homogeneity of the groups
The groups were equivalent regarding distribu-
tion of gender, stroke aetiology, affected side
of the body, age, body height and body mass.
816 T Wolny et al.

Table 1 Characteristics of the subjects with P-values of the homogeneity tests

Control group Exp 1 Exp 2 P-value

(n ¼ 32) (n ¼ 32) (n ¼ 32)

Females (%) 12 (51.85) 9 (28.12) 10 (51.85) 40.05*

Males (%) 20 (48.15) 23 (71.88) 22 (48.15)
Mean age 64,00 (7,33; 48–81) 59.87(8.34; 39–76) 61.47 (12.53; 32–82) 40.05**
(SD; min–max) years
Weight 79.97 (12.93; 77.31 (11.61; 76.12 (11.66; 40.05**
(SD; min–max) kg 44.00–112.00) 55.00–104.00) 52.00–105.00)
Height 169.37 (8.10; 168.50 (5.90; 170.56 (9.40; 40.05**
(SD; min–max) cm 152.00–187.00) 152.00–178.00) 154.00–194.00)
Post-stroke period 1.84 (0.78; 1.00–3.00) 1.73 (0.78; 0.00–3.00) 1.78 (0.75; 1.00–3.00) 40.05**
(SD; min–max) years
Left-side hemiplegia (%) 19 (59.37) 19 (59.37) 18 (56.25) 40.05*
Right-side hemiplegia (%) 13 (40.63) 13 (40.63) 14 (43.75)
Haemorrhagic stroke (%) 5 (15.62) 7 (21.87) 5 (15.62) 40.05*
Ischaemic stroke (%) 27 (84.38) 25 (78.13) 27 (84.38)

*Chi-squared test.
**One-way ANOVA.
Exp 1, experimental group 1; Exp 2, experimental group 2.

The characteristics of the three groups and out-
comes of homogeneity testing are presented in
Table 1.
Two-point discriminatory sense
With 32-subject groups we achieved statistical
power of ANOVA of 0.97 for two-point discrimi-
natory sense (power ¼ 1–beta (probability of com-
mitting the statistical error of the second kind).
Initial measurement showed a similar level of the
two-point discriminatory sense in the three groups
(P40.05 for all groups, Tukey test). Healthy limbs
were significantly better (P50.001 for all groups,
Tukey test). After the intervention, experimental
group 2 subjects were on average 4.17 mm better
than experimental group 1 subjects (95% CI 2.64
to 5.7) and 3.84 mm better than control group sub-
jects (95% CI 3.02 to 4.67). The affected limb out-
come in experimental group 2 was even 1.72 mm
(95% CI 0.20 to 3.23) better than for the healthy
limb (P40.05, Tukey test). Intra-group difference
for affected upper extremity was also significant
(P50.001, Tukey test). In experimental group 1
this difference was still negative, –3.88 mm (95%
CI –5.89 to –1.87) and significant (P50.001,
Tukey test) and so it was in control group (mean
3.22 mm; 95% CI –4.52 to –1.93; P50.001).
No significant inter-group differences were
registered. Detailed data on the two-point discrim-
inatory sense are presented in Table 2.
Stereognosia
The statistical power of ANOVA was 0.25 in
this case. A mean progression of 4.6 points (95%
CI 0.53 to 7.59) was noted in experimental group 1
and of 9.37 points (95% CI 1.24 to 17.51) in exper-
imental group 2 (P50.01, Tukey test) (Table 3).
In the control group we registered insignificant
deterioration of –0.62 points on average (95%
CI –1.90 to 0.65). No data on the healthy limb
were available for stereognosia. No significant
inter-group differences were registered. Because
of the low power of our statistical test a probabil-
ity of 75% of committing type 2 statistical error
should be considered here.
Thermaesthesia
In the case of thermaesthesia the statistical power
of ANOVA was 0.85. In the control group no ther-
apeutic effect was observed. In experimental group
1 one subject increased their score from 0 to 1 point.
The difference between the affected and the healthy
limb was always significant in these groups
(P50.001 for both the initial and final measure-
ments, Tukey test). In experimental group 2 the
mean difference between the affected and the
 Butler’s method in late-stage stroke subjects 817

Table 2 Mean  SD (min–max) of groups, mean (95% CI) inter-group differences and mean (95% CI) intra-group differences
for the two-point discriminatory sense (mm)

Measurement Control Exp 1 Exp 2 Inter-group

n ¼ 32 n ¼ 32 n ¼ 32
Control minus Control minus Exp 1 minus
Exp 1 Exp 2 Exp 2

ILI 4.56  1.50 4.22  1.23 5.66  4.88 0.34 –1.1 –1.44
(2.00–8.00) (2.40–7.80) (2.60–31.20) (0.23 to 0.43) (–2.31 to 0.11) (–2.74 to –0.12)
ALI 8.09  4.14 8.61  6.16 8.59  4.77 –0.52 –0.50 0.02
(2.00–18.60) (2.60–37.60) (3.60–26.40) (–1.25 to 0.20) (–0.73 to –0.28) (–0.48 to 0.52)
ALF 7.78  3.77 8.11  5.73 3.94  1.48 –0.33 3.84 4.17
(2.80–18.4) (2.40–34.60) (1.80–8.60) (–1.03 to 0.38) (3.02 to 4.67) (2.64 to 5.7)
Intra-group
ILI minus ALI –3.53* –4.38* –2.94*
(–4.92 to –2.14) (–6.56 to –2.20) (–4.10 to –1.77)
ILI minus ALF –3.22* –3.88* 1.72
(–4.52 to –1.93) (–5.89 to –1.87) (0.20 to 3.23)
ALF minus ALI –0.31 –0.50 –4.66*
(–0.57 to –0.04) (–0.86 to –0.13) (–6.06 to –3.25)

*P50.001, post-hoc Tukey test.

ILI, intact limb initial; ALI, affected limb initial; ALF, affected limb final; Exp 1, experimental group 1; Exp 2, experimental group 2.
Positive difference signifies bigger (worse) minuend, negative – bigger (worse) subtrahend.
No data on intact upper limb were registered in the final measurement.

Table 3 Mean  SD (min–max) of groups, mean (95% CI) inter-group differences and mean (95% CI) intra-group differences
for stereognosia [points]

Measurement Control Exp 1 Exp 2 Inter-group

n ¼ 32 n ¼ 32 n ¼ 32
Control minus Control minus Exp 1 minus
Exp 1 Exp 2 Exp 2

ALI 57.19  49.46 58.12  46.45 62.19  43.68 –0.93 –5.00 –4.07
(0–100) (0–100) (0–100) (–2.03 to 0.18) (–7.09 to –2.92) (–5.06 to –3.07)
ALF 56.56  50.07 62.19  45.49 71.56  41.43 –5.63 –15.00 –9.37
(0–100) (0–100) (0–100) (–7.28 to –3.98) (–18.11 to –11.89) (–10.83 to –7.91)
Intra-group
ALF minus ALI –0.62 4.06 9.37*
(–1.90 to 0.65) (0.53 to 7.59) (1.24 to 17.51)

*P50.01, post-hoc Tukey test.

ALI, affected limb initial; ALF, affected limb final; Exp 1, experimental group 1; Exp 2, experimental group 2.
Positive difference signifies bigger (better) minuend, negative – bigger (better) subtrahend.
No data on intact upper limb were registered in case of stereognosia.

healthy extremity was reduced from 0.56 points
(95% CI 0.25 to 0.88; P50.01, Tukey test) to
0.34 points (95% CI 0.13 to 0.56; P40.05, Tukey
test). No significant inter-group differences were
revealed (Table 4).
Change scores
Change scores for all parameters and all groups
may be found in the last rows of Tables 2, 3 and 4
(ALF minus ALI). In the case of the two-point
discriminatory sense experimental group 2 was
significantly better than the two other groups
(P50.001 in each case, Tukey test). A similar situ-
ation was registered for thermaesthesia (experimen-
tal 2 versus experimental 1 P50.01; experimental 2
versus control P50.001, Tukey test). For stereo-
gnosia the only significant difference was found
between experimental 2 and control group
(P50.05).
818 T Wolny et al.

Table 4 Mean  SD (min–max) of groups, mean (95% CI) inter-group differences and mean (95% CI) intra-group differences
for thermaesthesia (points)

Measurement Control Exp 1 Exp 2 Inter-group

n ¼ 32 n ¼ 32 n ¼ 32
Control minus Control minus Exp 1 minus
Exp 1 Exp 2 Exp 2

ILI 3.00  0.00 2.94  0.35 3.00  0.00 0.06 0.00 0.06
(3.00) (1.00–3.00) (3.00) (0.19 to 0.06) (–) (0.19 to 0.06)
ALI 2.31  0.10 2.06  1.16 2.44  0.88 0.25 –0.13 –0.38
(0.00–3.00) (0.00–3.00) (0.00–3.00) (0.19 to 0.31) (–0.17 to –0.08) (–0.48 to –0.27)
ALF 2.31  0.10 2.09  1.12 2.78  0.55 0.22 –0.47 –0.67(–0.89
(0.00–3.00) (0.00–3.00) (1.00–3.00) (0.17 to 0.26) (–0.62 to –0.31) to –0.48)
Intra-group
ILI minus ALI 0.69* 0.87* 0.56**
(0.33 to 1.05) (0.48 to 1.27) (0.25 to 0.88)
ILI minus ALF 0.69* 0.84* 0.22
(0.33 to 1.05) (0.44to 1.24) (0.02 to 0.42)
ALF minus ALI 0.00 0.03 0.34
(–) (–0.03 to 0.09) (0.13 to 0.56)

*P50.001, **P50.01 post-hoc Tukey test.

ILI, intact limb initial; ALI, affected limb initial; ALF, affected limb final; Exp 1, experimental group 1; Exp 2, experimental group 2.
Positive difference signifies bigger (better) minuend, negative – bigger (better) subtrahend.
No data on intact upper limb were registered in the final measurement.

Discussion progressed significantly better than the other

On first inspection of our results it may be thought
that the therapeutic protocols did not differ in
their effectiveness. Neither the pre-test nor the
post-test comparison of the results using mixed
ANOVA showed significant inter-group differ-
ences. However, a specific tendency was noticed
when comparing intra-group results. In experi-
mental group 2 we always observed the largest
improvement which was not present in other
groups. This improvement was expressed either
in the form of significant intra-group differences
(two-point discriminatory sense, stereognosia;
Tables 1 and 2) or reduction of differences with
the intact upper extremity (two-point discrimina-
tory sense, thermaesthesia; Tables 1 and 3).
In the case of the two-point discriminatory sense
the post-test demonstrated better outcome on the
affected side of the body than on the healthy one.
In such situations we decided to count and com-
pare change scores for all parameters. Here, the
results undoubtedly indicate the greater therapeu-
tic impact of the treatment using neuromobiliza-
tion. In the case of two-point discriminatory sense
and thermaesthesia experimental group 2
groups. In the case of stereognosia it was better
than the control group. In this case the progress
was double that of experimental group 1, but we
were unable to find statistical significance with
ANOVA and Tukey test (in the case of stereogno-
sia the reader should be aware of the low statistical
power of ANOVA and the high probability of
accepting a false null hypothesis).
Such a situation has certain clinical implica-
tions. It shows that upper extremity proprioceptive
neuromuscular facilitation patterns and mobi-
lizations of the peripheral nerves, although incor-
porating similar motions and positions (e.g.
flexion-abduction-external rotation pattern –
mobilization of the median nerve; extension-
abduction-internal rotation pattern – mobilization
of the radial nerve), should not be treated as alter-
native methods of treatment aiming to reduce sen-
sory deficits. It seems that more effective
neuromobilization is provided not only by the spe-
cific position of the limb, but also by rhythmic,
repeated, short increases of tension (about 1 Hz)
which are the essence of these techniques.
Despite of such interesting results we must also
emphasize the weaknesses of the study. First,

although the size of our sample was sufficient in
terms of statistics, it may well be too small to draw
definite clinical conclusions. Similarly, the dura-
tion of the treatment may seem a bit too
short for some clinicians. We also did not orga-
nize a longer follow-up, so we cannot be sure
whether any effects were temporary, long-term or
permanent.
We would also like to draw attention to some
problems that may arise during the application of
these techniques. In the case of neuromobilization
no standard, validated protocols aiming to obtain
the best outcome have been developed yet (dura-
tion, number of repetitions and sets, amount of
force, etc.). At present much depends on the
experience of the individual clinician. In the case
of post-stroke therapy this may be even more
complicated since sensory deficit may make
patient feedback to the clinician limited and
cause difficulties in adjusting the amount of force
appropriately.
Information about the effectiveness of Butler’s
neuromobilization applied as an alternative form
of stimulation supporting regular post-stroke
therapy is sparse. Earlier, Davies applied neuro-
mobilization in stroke patients31,32; he provided
technical details of the applied techniques, reported
a favourable influence on muscle tone and improve-
ment of some indicators of functional capacity, but
did not mention any specific objective tools aiming
to evaluate the outcome. Rolf33 provides some
information on neural tissue mobilization in
stroke patients and claims its positive effects; classic
Butler techniques were developed by this author to
meet the demands of the neurological patient.
Impaired neuromechanics is also addressed by the
proprioceptive neuromuscular facilitation method,
which allows the modification or combination of
therapeutic techniques so that they can be used
for peripheral nerve mobilization.34
Independent of their aetiology, hemiplegia and
hemiparesis of cerebral origin are a specific disor-
der of the locomotory system.35 Injury affecting
the central nervous system usually has a circula-
tory background, but major dysfunctions involve
the locomotory system. Normal neuromechanics
may be affected by changes located either within
the borders of the nervous system or outside it.
Paralysis or paresis spreading across one side of
the body, spastic muscle tone, pathologic
Butler’s method in late-stage stroke subjects 819
synergisms, motor impairment – all these symp-
toms may result in unfavourable muscle, connec-
tive and neural tissue changes, which not only
destroy their structure but also internal plasticity.
Pathological damage can arise because of the
reduced ability of the nerve to slide on neighbour-
ing tissue.
It should be noted that the peripheral nerve is a
mixed nerve containing sensory, motor and auto-
nomic fibres. All peripheral nerves provide the
central nervous system with various data concern-
ing exteroception, proprioception, thermaesthesia,
stereognosia, joint position sense, etc. This infor-
mation will only be appropriate if the function of
the nerve remains totally unspoiled. From this
point of view it seems that many stroke patients
may suffer from disturbed neuromechanics which
is followed either by impaired data transmission or
by transmission of misleading information.
Lasting pressure exerted on the neuron and on
surrounding connective tissue is able to introduce
fibrosis followed by distorted data transmission,
axonal transport, vasomotor and trophic distur-
bances. From this point on a short path leads to
blood supply impairment which affects inter-tissue
chemical reactions and finally reduces the overall
physical capacity of the patient. The nerve may
also receive confusing information from so-called
‘nervi nervorum’ and therefore become either a
source of its own symptoms or a factor cementing
pathological motion patterns and/or malalignment
of body segments.14
The important problem of affected exterocep-
tion, especially within the paralysed upper limb, is
frequently neglected. It is well known that a major
function of the hand (besides manual motor tasks)
is reception of external tactile information, ther-
maesthesia and stereognosia. Biocybernetic
models clearly indicate that only an adequate
input information to the control system allows
for an appropriate response. Therefore, modula-
tion of tension within all structures of the upper
extremity may provide the central nervous system
with desired information, which in turn allows for
adequate motor performance. Impairments of this
system may burden the upper limb with severe dis-
ability, with modification of the motor schemes
and deregulation of spatial and time parameters
of the movement in the background.
820 T Wolny et al.
All arguments presented above indicate that
post-stroke motor disturbances include dysfunc-
tion of the peripheral nerves. Therefore, it may
well be that mechanical stimulation of these
nerves activates certain therapeutic potential,
which is all too often neglected. Although our
results must be treated with caution and need fur-
ther investigation and analyses, they indicate –
according to our point of view – that application
of peripheral nerve mobilization is to be recom-
mended. These therapeutic techniques should
find their place in the therapeutic armamentarium
of every physiotherapist routinely dealing with
neurological problems connected with disturbed
neuromechanics.
Clinical messages
 A complex physiotherapy in which proprio-
ceptive neuromuscular facilitation patterns
are supported by mobilization of the periph-
eral nerves significantly improves two-point
discriminatory tactile perception, stereogno-
sia and thermaesthesia in late-stage post-
stroke patients.
 In situations in which it is impossible to
apply neuromobilization, the proprioceptive
neuromuscular facilitation treatment alone
may provide certain therapeutic benefits.
 Traditional post-stroke therapy did not
prove its impact on tactile perception,
stereognosia and thermaesthesia in patients
about 1–2 years after stroke.
References
1 Bobath B. Adult hemiplegia. Oxford: Heinemann
Medical, 1990.
2 Laidler P. Rehabilitacja po udarze mózgu (Post
stroke rehabilitation). Warsaw: PZWL, 2000.
3 Jakimowicz W. Neurologia kliniczna w zarysie
(Outline of clinical neurology). Warsaw: PZWL,
1987.
4 Winward CE, Halligan PW, Wade DT. Current
practice and clinical relevance of somatosensory
assessment after stroke. Clin Rehabil 1999; 13: 48–55.
5 Gola˜ b BK. Anatomia czynnos´ciowa os´rodkowego
ukiadu nerwowego (Functional anatomy of the ner-
vous system). Warsaw: PZWL, 1992.
6 Knott M, Voss DE. Proprioceptive neuromuscular
facilitation; patterns and techniques, second edi-
tion. New York: Harper and Row, 1968.
7 Rose L, Bakal DA, Fung TS, Farn P,
Weaver LE. Tactile extinction and function status
after stroke: a preliminary investigation. Stroke
1994; 25: 1973–6.
8 Kim JS, Choi-Kwon S. Discriminative sensory
dysfunction after unilateral stroke. Stroke 1996;
27: 677–82.
9 Cormon A, Benton AL. Tactile perception of
direction and number in patients with unilateral
cerebral disease. Neurology 1969; 19: 525–32.
10 Robertson SL, Jonnes LA. Tactile sensory impair-
ments and prehensile function in subjects with
left-hemisphere cerebral lesions. Arch Phys Med
Rehabil 1994; 75: 1108–17.
11 Samuelsson M, Samuelsson L, Lindell D. Sensory
symptoms and signs and results of quantitative
sensory thermal testing in patients with lacunar
infarct syndromes. Stroke 1994; 25: 2165–70.
12 Schabrun SM, Hillier S. Evidence for the retrain-
ing of sensation after stroke: a systematic review.
Clin Rehabil 2009; 23: 27.
13 Carey LM. Somatosensory loss after stroke. Crit
Rev Phys Rehabil Med 1995; 7: 51–91.
14 Butler DS. Mobilization of the nervous system.
Edinburgh: Churchill Livingstone, 1991.
15 Shacklock M. Clinical neurodynamics. Edinburgh:
Churchill Livingstone, 2005.
16 Nowotny J. Zarys rehabilitacji w dysfunkcjach
narza˜ du ruchu (The outline of rehabilitation in
locomotory dysfunctions). Katowice: AWF, 1990.
17 Mika T. Fizykoterapia (Physicotherapy). Warsaw:
PZWL, 1996.
18 Straburzyński G. Fizjoterapia (Physiotherapy).
Warsaw: PZWL, 1988.
19 Straburzyński G, Straburzyńska-Lupa A.
Medycyna fizykalna (Physical Medicine). Warsaw:
PZWL, 1997.
20 Adler S, Deckers D, Buck M. Proprioceptive neu-
romuscular facilitation in practise. Berlin: Springer
Verlag, 1993.
21 Kokosz M, Saulicz E, Z_ mudzka-Wilczek E,
Saulicz M. Mo_zliwość wykorzystania stymulacji
oraz technik specjalnych w metodzie PNF (The
possibility of use stimulation and specific tech-
niques in PNF method). Fizjoterapia 1998; 6: 3.
22 Voss DE, Ionta MK, Myers BJ. Proprioceptive
neuromuscular facilitation patterns and techniques.
Philadelphia: Harper and Row, 1985.
23 Kaltenborn F, Wirbelsäule. Manuelle
Untersuchung und Mobilization (Manual assess-
ment and mobilization). Oslo: Olaf Norlis
Bokhandel, 1995.

24 Szprynger J, Sozańska G. Neuromechanika i neu-
romobilizacje w fizjoterapii (Neuromechanics and
neuromobilizations in physiotherapy). Lublin:
Wydawnictwo Czelaj, 2001.
25 Crosby PM, Dellon AL. Comparison of two-
point discrimination testing devices. Micro Surg
1989; 10: 134–7.
26 Dellon AL. Evaluation of sensibility and re-educa-
tion of sensation in the hand. Baltimore: Williams
and Wilkins, 1981.
27 Dellon AL, Mackinnon SE, Crosby PM.
Reliability of two-point discrimination testing.
J Hand Surg 1987; 12A: 693–6.
28 Mackinnon SE, Dellon AL. Two-point discrimi-
nation tester. J. Hand Surg 1985; 10A: 906–7.
29 Knapik H. Zjawisko asymetrii funkcji kończyn u
chorych z niedowiadem poiowiczym w procesie
rehabilitacji (Functional asymmetry of the limbs
in hemiplegic patients). Wydawnictwo
Monograficzne no. 31. Krakow: AWF, 1988.
30 Pa˜ chalska M. Kompleksowy model rehabilitacji
chorych z ogniskowym uszkodzeniem mózgu i
Butler’s method in late-stage stroke subjects 821
afazja˜ caikowita˜ (Complex model of rehabilitation
in patients with focal brain injury and aphasia).
Wydawnictwo Monograficzne No. 29. Kraków:
AWF, 1986.
31 Davies P. Starting again. Early rehabilitation after
traumatic brain injury or other severe brain lesions.
Heidelberg: Springer, 1994.
32 Davies P. Steps to follow. The comprehensive treat-
ment of patients with hemiplegia. Berlin: Springer
Verlag, 2000.
33 Rolf G. The puzzle of pain, loss of mobility, eva-
sive movements and the self-management. Danske
Fysioterapeuter 2001.
34 Hartman K. Neurodynamiczność jako kombi-
nacja technik terapeutycznych Proprioceptive
Neuromuscular Facilitation (Neurodynamics as
combination of Proprioceptive Neuromuscular
Facilitation therapeutic techniques). Krakow:
Materialy z kursu, 2003.
35 Czlonkowska A, Czlonkowski A. Diagnostyka i
leczenie w neurologii (Diagnostics and treatment
in neurology). Warsaw: IPiN, 1992.