Clinical Biomechanics 16 (2001) 481±488

www.elsevier.com/locate/clinbiomech

Kinematics of rotational mobilisation of the lumbar spine

Raymond Y.W. Lee 1
Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Yuk Choi Road, Hunghom, Hong Kong
Received 1 November 2000; accepted 11 April 2001

Abstract
Objective. The purposes of this study were to measure the movements of the lumbar spine produced by rotational mobilisation,
and to study the e€ects of di€erent grades of mobilisation on the movements produced.
Design. Kinematics of rotational mobilisation was assessed with an electromagnetic tracking device.
Background. Rotational mobilisation is frequently used in the treatment of back pain, but there was no information on its
mechanical e€ects.
Methods. Movements of the lumbar spine were measured in 14 healthy volunteers when they were subjected to grades I to IV left
rotational mobilisation.
Results. In the starting positions, the spines were found to be ¯exed, axially rotated to the left and laterally bent to the right. As
the mobilisation grade increased, the spine was axially rotated further into the range. Rotational mobilisation was found to produce
oscillatory movements of the lumbar spine in all three anatomical planes. It produced axial rotation which was accompanied by
lateral bending in the opposite direction and sagittal rotation. The mean frequency of the oscillatory movements was 1.4 Hz. The
amplitude of the oscillations was small, and was found to be increased in grades II and III mobilisation.
Conclusion. Rotational mobilisation may be able to restore lost movements of the lumbar spine in any of the three anatomical
planes.

Relevance
An understanding of the kinematics of mobilisation will allow therapists to gain insight into its mechanical and perhaps ther-
apeutic e€ects. Ó 2001 Elsevier Science Ltd. All rights reserved.

Keywords: Manipulative therapy; Mobilisation; Manipulation; Biomechanics; Kinematics; Lumbar spine

1. Introduction largely con®ned to the sagittal plane of the lumbar

Spinal mobilisation is frequently employed by phys-
iotherapists and other clinicians in the treatment of low
back pain [1]. The evidence-base for this form of therapy
has increased considerably in recent years. For acute
and sub-acute back pain, mobilisation and manipulation
were found to provide better short-term improvement in
pain and activity levels and higher patient satisfaction
when compared to other conservative treatments [2±9].
The mechanical mechanisms underlying mobilisation
were studied by a number of research studies [10±14].
They had only examined posteroanterior mobilisation,
and two-dimensional measurements were made as the
loads applied and the resulting spinal motions were
E-mail address: rsrlee@polyu.edu.hk (R.Y.W. Lee).
1
Dr. Lee was a Senior Lecturer in the School of Physiotherapy, the
University of Sydney at the time the study was conducted.
0268-0033/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 6 8 - 0 0 3 3 ( 0 1 ) 0 0 0 3 6 - 5
spine. However, most other manual therapy techniques,
including the highly popular rotational mobilisation,
involve complex three-dimensional movements and
loading. The biomechanical characteristics of these
techniques had never been examined in previous re-
search. This had prompted the present study to examine
the three-dimensional kinematics of rotational mobili-
sation. Such knowledge is clinically important as it will
allow clinicians to understand the mechanical mecha-
nisms of actions and to develop strategies for restoring
movements which may be lost as a result of spinal pain
or pathologies.
Maitland [15] depicted that rotational mobilisation
produced axial rotation of the lumbar spine, and that
the axis of rotation was underneath the couch on which
the patient was lying. However, this was no experi-
mental evidence supporting his hypotheses. In fact, it
was demonstrated that axial rotation of the lumbar
482 R.Y.W. Lee / Clinical Biomechanics 16 (2001) 481±488
spine was inherently coupled with movements in other
planes [16±19]. Rotational mobilisation would thus elicit
a complex movement pattern rather than simple axial
rotation.
Clinically, rotational mobilisation may be performed
with di€erent ``grades''. This involves positioning the
patient in di€erent postures and twisting the trunk to
di€erent degrees [15]. It was demonstrated that spinal
posture had signi®cant in¯uences on the movements of
the spine [20], and thus it is possible that the grade of
mobilisation may have an e€ect on the movements
produced. However, clinically, the choice of mobilisa-
tion grade is largely empirical and there is no experi-
mental data on the e€ects of mobilisation grade on the
kinematics of the spine.
Measurement of spinal kinematics in three dimen-
sions has been fraught with diculties due primarily to
the inaccessibility of the spine. Radiographic measure-
ments [13,19,21] were commonly used, but it was pos-
sible only to obtain static measurements of the end
range of motions. Some previous researchers [22±25]
inserted wires or pins percutaneously into the spinous
processes, and measured the intervertebral movements
with protractors or movement sensors attached to the
pins. These invasive techniques were accurate, but they
might cause discomfort to the subjects limiting their use
to small groups of volunteers. Video or opto-electronic
techniques using surface markers had also been em-
ployed in studying spinal kinematics [26], but they were
generally time-consuming, expensive and inaccurate due
to closely positioned markers and changing inter-marker
distances.
Recently, electromagnetic tracking devices have been
used for kinematic studies of the spine [27±31]. They
would be most ideal for the present study as they are
non-invasive and would provide three-dimensional ki-
nematic information in real time. Previous studies
[28,30] showed that these devices were highly reliable for
measuring spinal movements. The data provided was
also found to have good correlation with measurements
using high precision potentiometers and radiographs
[31].
Clearly, there is a strong need to examine the three-
dimensional kinematics of rotational mobilisation. The
purposes of this study were to measure the spinal
movements produced by rotational mobilisation, and to
study the e€ects of di€erent grades of mobilisation on
the movements produced.
2. Methods
2.1. Subjects
Fourteen subjects (8 men and 6 women, mean
age ˆ 28 (SD ˆ 6), mean height ˆ 1.71 m (SD ˆ 0.17),
mean weight ˆ 65.6 kg (SD ˆ 12.7)) agreed to partici-
pate in this study. They were in good health with no
history of back pain or leg pain that may be attributed
to the back within the last 12 months. They were ex-
cluded if they had undergone previous back surgery, had
a fracture, dislocation or any structural defects of ver-
tebral structures.
The study was approved by the Ethics Committee of
the University of Sydney. Subjects were informed about
the experimental procedure and any potential risks prior
to the attainment of a written consent form.
2.2. Instrumentation
The 3SPACE Fastrak (Polhemus, Colchester, VT,
USA) was used to measure movements of the lumbar
spine produced by rotational mobilisation. The system
had a source that generated a low-frequency magnetic
®eld which was detected by the sensors. The source was
placed in a ®xed position close to the subject (within 0.7
m). Two sensors of the system were employed to mea-
sure the movements of the lumbar spine. One sensor was
placed over the L1 spinous process and the second
sensor over the sacrum. Each sensor was attached to a
small, mouldable plastic plate with plastic screws. A
Velcro band was threaded through the plate and tightly
wrapped around the subject's body so as to minimise the
movement between the sensor and the underlying skin.
The cables of the sensors were attached to the skin on
the side of the trunk so that they did not move the
sensors erroneously during mobilisation. Initial testing
showed that the above arrangement provided the most
secure sensor attachments.
The Fastrak had an electronic unit that calculated the
three-dimensional positions and orientations of the
sensors relative to the source. The unit was linked to a
personal computer via an RS232 serial interface. Spe-
ci®cally developed custom software was used to control
the Fastrak operation, data acquisition and display in
real time.
There were several improvements in the present
measurement system upon the electromagnetic devices
reported previously [28±31]. Firstly, the source was not
attached to the skin of the sacral area as in previous
studies [28±30]. This was possible as the Fastrak had
more than one sensor. Low-frequency vibration of the
source would be a problem if it were attached to the
sacrum [30]. Secondly, in previous studies [27±29,31],
the data update rate of the electromagnetic devices was
low (typically 20 Hz) due to the limited Baud rate. The
software developed in this study was able to perform
fast serial communication at 115.2 kBaud allowing a
data update rate of 120 Hz. As two sensors were used in
this experiment, the sampling rate was 60 Hz per sensor.
Finally, preliminary testing showed that the data update
rate was slightly inconsistent if the internal clock of the
 R.Y.W. Lee / Clinical Biomechanics 16 (2001) 481±488
Fastrak was used. There were variations in the order of
microseconds in each data update cycle. Thus in the
present study, the Fastrak was externally driven by
di€erential transistor transistor logic (TTL) compatible
pulses of 1 ms and 2:5 V to achieve an accurate update
rate.
The Fastrak output comprised the 3  3 matrices of
direction cosines that described the orientations of the
sensors relative to the source. The relative orientation
between the L1 and sacral sensors, which described the
movements of the lumbar spine, was derived from these
matrices. The anatomical angles of ¯exion±extension,
lateral bending and axial rotation were computed and
displayed on the computer screen in real time. The
method of computation was based on the mathematical
techniques proposed by earlier authors [32±34], and is
explained in Appendix A.
A pilot study was conducted to examine the reliability
of the data provided by the present measurement sys-
tem. The initial starting positions and the amplitude of
movements of the lumbar spines of eight normal sub-
jects were measured three times while grade III left ro-
tational mobilisation was performed. Statistical analysis
revealed excellent reliability for the data recorded in the
three measurement trials, with intraclass correlation
coecients ranging from 0.93 to 0.99. Errors in mea-
suring the positions or movements were de®ned as the
standard deviations of the measurements. The mean
measurement errors were found to be 1.4° for the initial
positions of the spine and 0.9° for the amplitude of
movements during mobilisation. It was concluded that
the data provided by the system was suciently reliable
for the purpose of this study.
2.3. Procedure
Prior to the study, subjects were requested to stand
upright with feet shoulder-width apart and palms facing
inwards. The position of the spine in this posture was
recorded by the Fastrak. The upright standing posture
was taken as the zero reference position, and the ana-
tomical angles were calculated with respect to this ref-
erence position (see Appendix A).
Left rotational mobilisation was carried out in this
study using the technique described by Maitland [15].
With grades I and II, the initial starting position adop-
ted was side-lying with ¯exion of both hips and knees.
Rotational mobilisation was then performed by pushing
the pelvis in an oscillatory manner. Small oscillatory
movements were used for grade I and large movements
for grade II. With grade III, the spine was twisted fur-
ther into the range by rotating the thorax to the left.
With grade IV, additional twisting force was achieved
by hanging the top leg over the edge of the couch. For
the stronger grades of III and IV, counter-pressure was
applied to the left shoulder while the pelvis was pushed.
483
Large oscillatory movements were used for grade III
and small movements for grade IV as advocated by
Maitland. Therapists who participated in this study had
received previous training in Maitland's technique of
rotational mobilisation, and had at least ®ve years of
clinical experience in musculoskeletal physiotherapy.
They were instructed to carry out the mobilisation using
what they would normally use in the clinical situation,
and the grade of mobilisation was judged using their
clinical experience. Di€erent grades of mobilisation were
performed by the same therapist for each subject in a
random sequence.
The initial starting position of the spine was recorded
for each grade of mobilisation before any mobilisation
was performed. Data was not collected in the ®rst
minute of mobilisation. Initial testing showed that
movements of the spine were not stable during this pe-
riod of time as the therapist was accommodating to the
resistance of the spine. Data was collected over 8 s after
the ®rst minute, and thereafter, the mobilisation was
terminated.
2.4. Statistical analysis
Repeated measures one-way analysis of variance
(A N O V A ) was performed using SPSS Version 10.0.0 to
study the e€ects of mobilisation grade on the dependent
variables ± initial starting positions of the spine and the
movements of the spine during mobilisation. Repeated
measures A N O V A was used because the same subjects
were tested under di€erent conditions (di€erent mobili-
sation grades). Using a standard A N O V A in this case was
not appropriate because the data would violate the
A N O V A assumption of independence. If the e€ect of
mobilisation grade was found to be signi®cant on a
dependent variable, post-hoc analysis was carried out
using the Tukey procedure. Statistical signi®cance was
accepted at the 5% level.
3. Results
3.1. Initial starting positions
Fig. 1 shows the mean posture of the lumbar spine in
the initial starting positions for di€erent grades of left
rotational mobilisation. In the starting positions, the
spines were found to be ¯exed, axially rotated to the left
and laterally bent to the right. As the mobilisation grade
increased, the spine was axially rotated further into the
range, from a mean value of 5.3° for grade I to 12.1° for
grade IV …P < 0:05†. Post-hoc analysis showed that the
amount of axial rotation for the four mobilisation
grades was signi®cantly di€erent from each other
…P < 0:05†. However, A N O V A revealed that increasing
the mobilisation grade did not signi®cantly alter the
484 R.Y.W. Lee / Clinical Biomechanics 16 (2001) 481±488
Fig. 1. Mean posture of the lumbar spine in the initial starting posi-
tions for di€erent grades of left rotational mobilisation.
posture of the spine in the sagittal and coronal planes
…P > 0:05†.
3.2. Movements of the lumbar spine produced during
mobilisation
Typical movement patterns for one of the subjects are
presented in Fig. 2. Mobilisation was found to produce
oscillatory movements in all three anatomical planes.
Axial rotation was approximately half a cycle out of
phase with movements in the other planes, whereas
movements in the sagittal and coronal planes were ap-
proximately in phase. In other words, when the spine
was mobilised into left axial rotation, it was also ex-
Fig. 2. Movements of the lumbar spine during grade II left rotational
mobilisation in one of the subjects (female, 23 yr old).
tended and laterally bent to the right. The opposite
movement pattern occurred in the next half of the mo-
bilisation cycle. All subjects showed the above move-
ment pattern, and the grade of mobilisation was found
not to alter the pattern.
In order to determine the frequency and amplitude of
mobilisation from the movement-time histories (Fig. 2),
data was ®tted with the sinusoidal function
Y ˆ A sin…2pft ‡ P † ‡ C;
where Y is the observed spinal movement, A is a con-
stant (the maximum magnitude of oscillation from its
averaged centre), f is the frequency of oscillation, C is
the average centre of the oscillation, and P is the phase
shift. The curve ®tting was performed using the least-
squares method with the Davidon±Fletcher±Powell al-
gorithm [35]. In all cases, R was found to be larger than
0.92 showing a good degree of ®t. The mean frequency
of mobilisation was 1:4  0:2 Hz. The mean peak-to-
peak amplitude of the oscillatory movements is shown in
Fig. 3. The amplitude was generally small, with mean
amplitude ranging from 0.8 to 2.6°. Mobilisation grade
had signi®cant e€ects on the amplitude of sagittal ro-
tation, axial rotation and lateral bending …P < 0:05†.
Post-hoc analysis showed the amplitude of grades II and
III mobilisation was larger than that of grades I and II
for movements in all three anatomical planes
…P < 0:05†.
4. Discussion
This paper reports the ®rst experimental study of the
three-dimensional kinematics of rotational mobilisation
of the lumbar spine. The movement patterns observed
were consistent in all subjects, enabling conclusions to
be drawn on the mechanical e€ects of mobilisation. This
study was conducted in a group of young healthy sub-
jects, and would provide useful normative data for
comparison in future studies.
Fig. 3. Mean peak-to-peak amplitude of the oscillatory movements of
the lumbar spine for di€erent grades of mobilisation.
 R.Y.W. Lee / Clinical Biomechanics 16 (2001) 481±488
In the initial starting position, the lumbar spine was
positioned in some degree of ¯exion, with mean values
ranging from 26.0 to 28.8° (see Fig. 1). This was about
half of the maximum range of ¯exion of the lumbar
spine as reported in previous studies [19,28,31]. The
¯exed posture was achieved by ¯exing the hips and
knees as suggested by Maitland [15]. In the coronal
plane, the spine was only slightly bent with mean values
ranging from 4.7 to 5.4° (see Fig. 1). This was only
about one-®fth of the maximum range of lateral bending
[19,28,31].
It was shown that the spine was positioned in dif-
ferent amount of axial rotation for di€erent mobilisa-
tion grades. For grades I and II mobilisation, the
amount of axial rotation was small. However, for
grade IV mobilisation, the mean movement was 12.1°,
which was close to the maximum range of axial rota-
tion reported elsewhere [19,28,31]. The above ®ndings
suggest that the Maitland technique is e€ective in po-
sitioning the spine further in the inner range of axial
rotation as the mobilisation grade increases, and with
grade IV, the spine is positioned almost at its physio-
logical limit.
It may be deduced that for grades I and II mobili-
sation, the posterior annulus and ligaments will be
subjected to some tensile strain with the spine in mid-
¯exion [36]. For grade IV mobilisation, as the lumbar
spine is axially twisted to the end of range in a mid-
¯exion posture, there may be marked increase in the
loads on the compression facets and the strains of the
posterior and posterolateral annular ®bres [37]. How-
ever, it remains to be seen whether the above deductions
are true, and whether the mechanical changes produce
any therapeutic e€ects.
This study shows that rotational mobilisation does
not simply produce axial rotation of the lumbar spine. It
elicits a movement pattern which is more complex than
the name may imply, and involves movements in all
three anatomical planes. It was demonstrated that axial
rotation was always accompanied by lateral bending in
the opposite direction. This ®nding is in agreement with
the observation reported in previous studies [28,31]. The
mean frequency of the oscillatory movements was found
to be 1.4 Hz. This is similar to the frequency that was
reported in previous studies involving other types of
mobilisation techniques [10±12]. The mean amplitude of
the movements was generally small, in terms of a few
degrees. The higher amplitude observed for grades II
and III movements is in accordance with the de®nitions
of Maitland [15].
The gross movement perceived by the therapist dur-
ing rotational mobilisation will include movements of
the thoracic and lumbar spine, the pelvis and the hip.
The oscillatory movements at the lumbar spine are small
in amplitude, and contribute to the gross movements to
a small extent only. Hence, the resistance perceived by
485
the therapist may not necessarily represent the sti€ness
of the lumbar spine. Movements of the back must be
measured during therapy if clinicians would like to have
a knowledge of the mechanical e€ects that they have
actually produced.
The results of the present study show that rotational
mobilisation has the potential of restoring axial rotation
of the lumbar spine which may be lost as a result of
mechanical disorders of the back. Increasing the mo-
bilisation grade may help regain axial rotation in the
inner range. In addition, therapists should be aware of
the fact that rotational mobilisation produces three-di-
mensional mechanical e€ects, and it may also be used to
regain lateral bending and sagittal rotation.
There were several improvements in the electro-
magnetic tracking technique of the present study when
compared to those used in previous research [28±31].
The source was ®xed to the ground rather than the
body eliminating any errors due to low frequency vi-
bration of the source. The present system was also
capable of collecting data with an accurate sampling
rate which was high enough for movement analysis.
Since the system is relatively inexpensive, highly por-
table and easy to use, it is possible that it can be used in
routine clinical measurements. Perhaps the major at-
traction of the present system is that it provides real-
time information on how the spine moves during
therapy. This is extremely valuable in providing im-
mediate feedback to students learning manipulative
therapy, and to therapists requiring immediate clinical
evaluation.
The present measurement system was suciently re-
liable for the purpose of this study. However, any
measurement system that attempts to quantify lumbar
movements by measurement of movements of sensors or
markers attached to the overlying skin will su€er from
the movement of soft tissues disguising the true verte-
bral movements. In order to minimise such an error,
mounting plates and Velcro straps were used to provide
the most secure sensor placements possible. The mea-
surement errors were found to be small, and represented
the accuracy that could be achieved by surface mea-
surements without any invasive procedures. However, it
was noted that the amplitude of grades I and IV mo-
bilisation was comparable to the magnitude of the
measurement errors in some subjects. In such cases,
caution should be exercised in the interpretation of data,
and the amplitude of movements could not be ascer-
tained with con®dence. Nevertheless, the same move-
ment patterns were consistently observed in all subjects
for all grades of mobilisation. Hence, there was sucient
evidence for conclusions to be drawn about the move-
ment patterns.
One limitation of the present study was that it did not
measure the loads acting at the lumbar spine. Previous
research [38] had measured the spinal loads during high-
486 R.Y.W. Lee / Clinical Biomechanics 16 (2001) 481±488

velocity manipulative thrust using a specially con-

structed couch which was mounted onto a force plate.
Future research may examine the loads during rota-
tional mobilisation using a similar technique. It is rec-
ommended that the loads and movements be measured
simultaneously so as to determine the load-deformation
characteristics and sti€ness of the spine. It should also
be pointed out that the results of the present study may
be applicable to young healthy subjects only. The
movement patterns of the spine in subjects with back
pain will likely be di€erent. Another recommendation of
future research will be to examine the e€ects of back
pain and spinal pathologies on the mechanical e€ects of
mobilisation.

5. Conclusion

This study showed that in the initial starting positions

of left rotational mobilisation, the spines were found to Fig. 4. Orthogonal base vector systems for the L1 vertebra and
sacrum.
be ¯exed, axially rotated to the left and laterally bent to
the right. The spines were twisted further into the inner 2 3 2 3
range of axial rotation as the mobilisation grade in- AL1x AL1y AL1z R11 R12 R13
creased. Mobilisation was found to produce oscillatory 6 7 6 7
6 AL2x AL2y AL2z 7 6 R23 7
movements in all three anatomical planes. It produced 4 5 ˆ 4 R21 R22 5
axial rotation of the spine which was accompanied by AL3x AL3y AL3z R31 R32 R33
lateral bending in the opposite direction and sagittal 2 3
AS1x AS1y AS1z
rotation. It is concluded that rotational mobilisation
6 7
may be able to restore lost movements of the lumbar 6 7
4 AS2x AS2y AS2z 5;
spine in any of the three anatomical planes. The am-
plitude of the oscillatory movements of the lumbar spine AS3x AS3y AS3z
was found to be small. Such movements do not con- or
tribute signi®cantly to the gross movements of the trunk
and pelvis induced during mobilisation. Future research ‰AL Š ˆ ‰RŠ ‰AS Š;
is recommended to examine the load-deformation 1 T
‰RŠ ˆ ‰AL Š ‰AS Š ˆ ‰AL Š ‰AS Š ;
characteristics of rotational mobilisation and the e€ects
of pain and pathologies on the mechanical e€ects of the where x, y and z are the axes of the global reference
technique. system (ground), and ‰AS ŠT is the transpose of ‰AS Š.
The anatomical base vector sets ‰AL Š and ‰AS Š are
unknown, but the Fastrak provides information on the
orientation or base vector sets of the sensors attached
Acknowledgements
to L1 and the sacrum ‰SL Š and ‰SS Š. Assuming that the
sensors are rigidly attached to the body, the anatom-
The author would like to acknowledge the ®nancial
ical and sensor axes will have ®xed spatial relation-
support of the University of Sydney Research Grant
ships.
Scheme.
‰SL Š ˆ ‰ML Š ‰AL Š;
‰SS Š ˆ ‰MS Š ‰AS Š;
Appendix A
where ‰ML Š and ‰MS Š are matrices which de®ned the
The relative orientation of the L1 vertebra and the spatial relationships.
sacrum represented the posture or movement of the If the upright standing posture is taken as the zero
lumbar spine. Such orientation may be described by a reference position, then
matrix ‰RŠ which expressed the orthogonal base vector ‰AL Šupright ˆ ‰AS Šupright ˆ ‰IŠ;
set of the L1 vertebra ‰AL Š in terms of that of the sacrum
‰AS Š (Fig. 4). where ‰IŠ is the unit matrix.
 R.Y.W. Lee / Clinical Biomechanics 16 (2001) 481±488
Thus,
‰ML Š ˆ ‰SL Šupright ;
‰MS Š ˆ ‰SS Šupright :
It is then possible to express ‰RŠ in terms of the base
vector sets of the sensors.
T
‰RŠ ˆ ‰AL Š ‰AS Š ;
T T
‰RŠ ˆ ‰SL Š ‰ML Š ‰SS Š ‰MS Š;
‰RŠ ˆ ‰SL Š ‰SL ŠTupright ‰SS ŠT ‰SS Šupright :
The three anatomical angles (a, b and c) can be com-
puted from ‰RŠ using the method of Grood and Suntay
[32]. It can be shown that
p torsion. J Biomech Eng 1979;101:46±52.
a ˆ cos 1 R13 ; [19] Pearcy MJ. Stereoradiography of lumbar spine motion. Acta
 2
 Orthop Scand Suppl 1985;212:1±45.
R23
b ˆ tan 1 ;
R33
 
R12
c ˆ tan 1 : Orthop Rel Res 1992;279:60±7.
R11
References
[1] Foster NE, Thompson KA, Baxter GD, Allen JM. Management
of non-speci®c low back pain by physiotherapists in Britain and
Ireland. A descriptive questionnaire of current clinical practice.
Spine 1999;24:1332±42.
[2] Glover JR, Morris JG, Khosla T. Back pain: a randomised
clinical trial of rotation manipulation of the trunk. Br J Ind Med
1974;31:59±64.
[3] Sims-Williams H, Jayson MIV, Young SMS, Baddeley J, Collins
E. Controlled clinical trials of mobilisation and manipulation for
patients with low back pain in general practice. Br Med J
1978;2:1338±40.
[4] Sims-Williams H, Jayson MIV, Young SMS, Baddeley J, Collins
E. Controlled clinical trials of mobilisation and manipulation for
low back pain: hospital patients. Br Med J 1979;2:1318±20.
[5] Hadler NM, Curtis P, Gillings DB, Stinnett S. A bene®t of spinal
manipulation as adjunctive therapy for acute low back pain: a
strati®ed controlled trial. Spine 1987;12:703±6.
[6] Meade TW, Dyer S, Brownek W, Frank AO. Randomised
comparison of chiropractic and hospital outpatient management
for low back pain: results from extended follow up. Br Med J
1995;311:349±51.
[7] Koes BW, Assendelft WJJ, Van der Heijden GJMG, Bouter LM.
Spinal manipulation and mobilization for low back pain: an
updated systematic review of randomized clinical trials. Spine
1996;21:2860±71.
[8] Royal College of General Practitioners, Clinical guidelines for the
management of acute low back pain. London: RCGP; 1996.
[9] Andersson GB, Lucente T, Davis AM, Kappler RE, Lipton JA,
Leurgans S. A comparison of osteopathic spinal manipulation
with standard care for patients with low back pain. N Engl J Med
1999;341:1426±31.
[10] R.Y.W. Lee, The biomechanical basis of spinal manual therapy.
Ph.D. thesis, the University of Strathclyde, Glasgow; 1995.
[11] Lee RYW, Evans JH. Load-displacement-time characteristics of
the spine under posteroanterior mobilisation. Aust J Physiother
1992;38:115±23.
487
[12] Lee RYW, Evans JH. Towards a better understanding of
posteroanterior mobilisation. Physiother 1994;80:68±73.
[13] Lee RYW, Evans JH. An in-vivo study of the intervertberal
movements produced by posteroanterior mobilisation. Clin
Biomech 1997;12:400±8.
[14] Lee RYW, Evans JH. The role of spinal tissue in resisting
posteroanterior forces applied to the lumbar spine. J Manipula-
tive Physiol Ther 2000;23:551±6.
[15] Maitland GD. Vertebral Manipulation. 5th ed. London: Balti-
more; 1986.
[16] Markolf KL. Deformation of the thoracolumbar intervertebral
joints in response to external loads: a biomechanical study using
autopsy material. J Bone Joint Surg [Am] 1972;54:511±33.
[17] Pope MH, Wilder DG, Matteri RE, Frymoyer JW. Experimental
measurements of vertebral motion under load. Orthop Clin North
Am 1977;8:155±67.
[18] Schultz AB, Warwick DN, Berkson MH, Nachemson AL.
Mechanical properties of human lumbar spine motion segments,
Part II, Responses in ¯exion, extension, lateral bending and
[20] Panjabi MM, Yamamoto I, Oxland T, Crisco J. How does posture
a€ect coupling in the lumbar spine. Spine 1989;14:1002±11.
[21] Pope MH, Frymoyer JW, Krag MH. Diagnosing instability. Clin
[22] Gregersen GG, Lucas DB. An in vivo study of the axial rotation
of the human thoracolumbar spine. J Bone Joint Surg [Am]
1967;49:247±62.
[23] Gunzburg R, Hutton W, Fraser R. Axial rotation of the lumbar
spine and the e€ect of ¯exion. An in vitro and in vivo
biomechanical study. Spine 1991;16:22±8.
[24] Kaigle AM, Pope MH, Fleming BC, Hansson T. A method for
the intravital measurement of interspinous kinematics. J Biomech
1992;25:451±6.
[25] Ste€en T, Rubin RK, Baramki HG, Antoniou J, Marchesi D,
Aebi M. A new technique for measuring lumbar segmental
motion in vivo. Method, accuracy, and preliminary results. Spine
1997;22:156±66.
[26] Pearcy MJ, Gill JM, Hindle RJ, Johnson GR. Measurement of
human back movements in three-dimensions by opto-electronic
devices. Clin Biomech 1992;2:199±204.
[27] An KN, Jacobsen MC, Berglund L, Chao EYS. Application of a
magnetic tracking device to kinesiologic studies. J Biomech
1988;21:613±20.
[28] Pearcy MJ, Hindle RJ. New method for the non-invasive three-
dimensional measurement of human back movment. Clin Bio-
mech 1989;4:73±9.
[29] Adams MA, Dolan P. A technique for quantifying the bending
moment acting on the lumbar spine in vivo. J Biomech
1991;24:117±26.
[30] Burnett AF, Barrett CJ, Marshall RN, Elliott BC, Day RE.
Three-dimensional measurement of lumbar spine kinematics for
fast bowlers in cricket. Clin Biomech 1998;13:574±83.
[31] Mannion A, Troke M. A comparision of two motion analysis
devices used in the measurement of lumbar spinal mobility. Clin
Biomech 1999;14:612±9.
[32] Grood ES, Suntay WJ. A joint coordinate system for the clinical
description of three-dimensional motions: application to the knee.
J Biomech Eng 1983;105:136±44.
[33] Pearcy MJ, Gill JM, Whittle MW, Johnson GR. Dynamic back
movement measured using a three-dimensional television system.
J Biomech 1987;20:943±9.
[34] Cole GK, Nigg BM, Ronsky JL, Yeadon MR. Application of the
joint coordinate system to three-dimensional joint attitude and
movement representation: a standardization proposal. J Biomech
Eng 1993;115:344±9.
488 R.Y.W. Lee / Clinical Biomechanics 16 (2001) 481±488

[35] Fletcher R. Practical methods of optimization, 2nd ed. New
York: Wiley; 2000.
[36] Shirazi-Adl A, Ahmed AM, Shrivastava SC. A ®nite element
study of a lumbar motion segment subjected to pure sagittal plane
moments. J Biomech 1986;19:331±50.
[37] Shirazi-Adl A, Ahmed AM, Shrivastava SC. Mechanical response
of a lumbar motion segment in axial torque alone and combined
with compression. Spine 1986;11:914±27.
[38] Triano J, Schultz AB. Loads transmitted during lumbosacral
spinal manipulative therapy. Spine 1997;22:1955±64.