Three-dimensional kinematics of the human back

Clinical Biomechenics 1990; 5: 218-228

Three-dimensional kinematics of the human back RJ Hindle PhD’ MJ Pearcy PhD, c ~ng’ AT Cross FRCS* DHT Miller FRCS~ ‘Centre for Biomedical Engineering, University of Durham, UK 2Sunderland District General Hospital, Sunderland, UK 3North Tees General Hospital, Stockton, UK

Summary This study examined the use of a new device, the ~SPACE ISOTRAK, to measure the kinematic behaviour of the backs of normal subjects and of patients from two orthopaedic practices. The device was shown to measure angles with a root mean square error of less than 0.2” and individuals showed a maximum standard deviation from the mean of movements repeated five times of less than 4”. Eighty normal subjects both male and female in four age ranges 20-29,30-39,40-49 and 50 years and over were measured. The three-dimensional kinematic patterns were similar for all groups, showing normals to have clearly identified patterns of movement. A general trend for decreasing mobility with age and some sex differences were also demonstrated. The clinical series showed widespread disruption to the primary and coupled movements of all the patients. When grouped together by clinical diagnosis the patient groups showed evidence of discrete and identifiable alterations from the normal kinematic patterns. However, further studies of homogeneous patient groups are required in order to demonstrate whether the measurements are of actual clinical use. The clinical studies also demonstrated that clinical subjective assessment of back mobility bore little relation to the true movements and therefore the clinical measurement of back movements should be reassessed. This study has shown the ~SPACE ISOTFIAKto be an effective tool in the clinic for the three-dimensional kinematic measurement of low back mobility. Relevance The accurate measurement of three-dimensional kinematics is essential to understand how the normal human back functions and, clinically, to determine if identification of abnormal movement patterns is of use to the clinician. Key words: Three-dimensional measurement, kinematics, human spine,

Measurement of human back movements is a routine part of clinical practice in the investigation of patients with back pain. The premise behind these measurements is that alterations to these movements are caused

Received: 7 March 1989 Accepred: 11 May 1990 Correspondence and reprint requests to: Dr MJ Pearcy, Department

of Orthopaedic Surgery and Trauma, Royal Adelaide Hospital, North Terrace, Adelaide, 5000, South Australia, Australia 0 1990 Butterworth-Heinemann (XX&-0033/90/040218-11

Ltd

~SPACE ISOTRAK.

by pathology and hence an assessment of changes in movements will provide information on a patient’s disorder. Most of the techniques used give one-dimensional ranges of movement from devices mounted on the surface of the patient’s back’. Although the clinician is able to record a number giving some index of the patient’s back movement there is often very little correlation between these measurements and true spinal movement, and little evidence that they give the clinician more information than subjective observations of restricted movement 2*3. However, although there is evidence that alterations to movements are related to pathology4-‘, there is no definitive evidence in the litera-

Hindle et al.: Kinematics of human back

ture as to the clinical value of accurate measurements of back movement. Most people will suffer from back pain during their livessq9 and although no more than 3 to 10% of cases result in medical consultation the numbers involved are still considerable”. The bioengineer is under obligation to assess measurement techniques used clinically and to develop new ones that provide the information required by the clinician to aid diagnosis. The unreliability of simple one- and two-dimensional range of movement measurements, together with their lack of correlation with true spinal movements, lays the onus on the bioengineer to provide the clinician with techniques to quantify back and spinal movements accurately. The spine is a complex three-dimensional structure and it should thus be expected to undergo complex movements in three-dimensions. Alterations to movements produced by a spinal disorder may not affect the total range of movement, but the manner or pattern of the movement. Ideally a measurement system should therefore measure the three-dimensional kinematic behaviour of the back or spine. The clinical measurement of back movements requires a technique that is non-invasive, cheap, simple to use, and not time consuming either for the collection of data or their analysis to produce a preseniation of movements meaningful to the clinician. Following the examination of several techniques for the three-dimensional measurement of intervertebral movements” and low back movements” a new device from the American aerospace industry, the ~SPACEISOTRAK, was introduced as a possible technique for the measurement of three-dimensional back movements in the clinic13. The aim of this study was to measure the three-dimensional kinematics of the lower back in a normal group and to assess whether patients being considered for surgery because of back disorders had identifiable differences in their patterns of movement. Methods 3SPACE

ISOTRAK

~SPACE ISOTRAK is an electro-magnetic device for the measurement of the position and orientation of a sensor in space (Figure 1). Krieg14*15introduced the system developed by the Polhemus Navigation Science Division of the McDonell Douglas Electronics Company (USA) and suggested that it should be evaluated by the research community as a feedback mechanism for paraplegics. An et a1.l6 suggested that the system could have applications for kinesiologic study, and prior to this Buchalter et al.” first suggested the device as a possible tool for measuring spinal motion and reported preliminary trials. More recently the same research group has examined the mobility of back pain patients’8,‘9 followed by a more detailed account of their technique and its application to a study of lumbar brace immobilization20*2’. However, they have only used the system to record in-

The

Figure 1.

~SPACE ISOTRAK

219

showing electronics box, source and

sensor.

dices of motion statically at the extremes of movement. To the best of our knowledge no one has yet published reports describing the use of the ISOTRAK to record kinematic movement of the spine. Resolution accuracy and repeatability The inherent error of’this system in measuring angles has

been presented previously13. ISOTRAK has a resolution of less than 0.1” and the accuracy in measuring known angles was shown to reduce as the angle increased according to the regression equation: y = 1-56x+0-509 where y = true angle and x = ISOTRAK reading. The repeatability of measurement was studied using a hinged wooden jig. These trials indicated that the total root mean square error encountered in measuring angles with the device was less than O-2”and the results showed that the accuracy and repeatability of measurement in each plane were not affected by rotations in other planes. Attachment of source and sensor The major problem with any non-invasive system, such

as this, is the attachment of the measurement devices to the subject and ensuring that once attached they record the actual movement of the spine. A system for the attachment of the source over the sacrum on a moulded plastic pad and the sensor over the L1 spinous process with double-sided tape and a strap around the trunk was described previously13. The technique used for indentifying the L1 spinous process was that of Burton22, who has recently questioned the techniques used to identify various spinal landmarks in other non-invasive measurement studies2j. He reports that most authors simply state that, for example, the spinous process of L1 was identified by palpation, whilst McConnell et a1.24have shown how even trained personnel have difficulty in correctly locating spinal segments. Burton identifies the spinous process of L4 as being at the bisection of a line joining the highest points of the iliac crests22, based on the earlier findings of MacGibbon and Farfan2’. Having identified the spinous process of

220

C/in. Biomech. 1990; 5: No 4

secured over the spinous process of Li. Preliminary trials with this arrangement produced satisfactory results when measuring sag&al plane movements, but on attempting to measure axial rotation deformation of the skin was found to have a significant effect. This skin movement was prevented by placing a strap over the sensor and around the trunk of the subject, thus maintaining the sensor over the Li spinous process even during axial rotation. Figure 2 shows the source and sensor in place on a subject. Movement measurement

Figure 2. Source and sensor attached to subject.

Ld in this manner the spinous process of Li is then found by counting up the spinous processes. In some subjects this was made easier by getting the subject to flex slightly, making the spinous processes more prominent. The sensor is relatively light and so it was possible to attach it directly to the skin with the use of double-sided tape. However, one of the major concerns with non-invasive studies such as this, is that the movement of the skin, and hence the sensor attached to it, may not reflect the actual intervertebral movement that is occurring underneath it. Stokes26 secured steel markers to the skin overlying lumbar spinous processes and then measured sagittal and lateral flexion radiographically and compared the movements of the skin markers to those of the underlying vertebrae. The markers were found to agree with intervertebral markers to within about 10%. No anatomical data could be found to clarify the attachment or non-attachment of skin to underlying spinous processes. However, in this study, where whole lumbar movement is being considered, as opposed to segmental mobility, the attachment of the sensor to the skin can certainly be considered sufficiently accurate to give a measure of lumbar movement. The most effective means of attaching the sensor to the skin was found to be the use of two strips of doublesided tape, approximately 25 mm long, attached to the sensor in the shape of a diagonal cross which was then

Clinical assessment of back movements usually consists of the patient performing forward, backward and side bending and twisting. To permit comparison with existing techniques these were the movements measured in all of the subjects and patients investigated in this study. Each subject performed three movements, starting and ending in the relaxed upright standing posture: maximal flexion and extension, lateral bend to left and right and axial rotation to left and right. These movements were repeated starting in the opposite direction (e.g. extension and flexion). Each movement was performed during a 10 s period and data were collected at a frequency of 10 Hz, this having been found adequate in previous studies for the measurement of these relatively slow movements’*~*‘. After the source and sensor had been attached, the subject would first practise the movements. He/she would then perform maximal flexion and extension, the start of the measurement period being signalled by an audible cue from the computer. As the subject performed the movement the operator would then count out the 10 s in order that the movement could be performed smoothly. Lateral bend was performed similarly with the subject attempting to stretch the appropriate hand down each leg as far as was possible. Axial rotation was performed with the subjects’ arms crossed over their chests. Using this technique the actual subject contact time could be as little as 10 min. The data were recorded using a lap-top computer which enabled the system to be completely portable. Movement repeatability The repeatability of an individual’s movements and their measurement with the ISOTRAK have been reported elsewherez. Three subjects each performed maximal voluntary flexion and extension, lateral bend and axial rotation five times in succession. They then performed extension and flexion, lateral bend, starting to the opposite side, and axial rotation, starting to the opposite side five times in succession. The standard deviation about the mean throughout the range of movement in each set of repeated movements ranged from O-39-3.92” with a mean of 1.84“. This value was judged to be sufficiently small that if a subject were to repeat a movement twice only, starting in

Hind/e et al.: Kinematics of human back

221

Table 1. Ages of normal subjects in years

Males

Females

20-29

30-39

40-49

2 50

20-29

30-39

40-49

3 50

: 3 4 5 6 7 8 9 10

22 26 25 27 28 26 20 26 29 27

37 35 36 32 32 35 32 38

53 55 65 59 57 58 53 64 55 59

23 21 21 ::

39 31 32 36 38 38 33 32

z:

43 40 49 45 45 41 44 42 40 40

El

46 44 44 41 43 42 45 46 43 40

50 54 52 54 53 56 57 53 51 51

Mean

26

35

43

58

25

34

43

53

n

opposite directions, obtained.

representative

results would be

Normal movements measurement

This section describes the collection of a data base of normal subjects prior to the measurement of patients. Subjects The movements of 80 individuals, 40 male and 40 female, with 10 in each of the four age ranges 20-29,3039,40-49 and 2 50 yr were recorded. Details of the ages of subjects are given in Table 1. None of these subjects had experienced any low back pain during the previous six months and none had ever had surgery on their backs. Procedure, data analysis and statistics The basic procedure for measurement remained the same as that described above. After being given an explanation as to the nature of the study and what was going to be required of him or her, subjects performed six movements: flexion and extension, extension and flexion, lateral bend to both sides, starting first to the left and then to the right and axial rotation to both right and left, starting to each side in turn. During the procedure the subjects were exhorted to bend or twist as far as they could. Every individual performs movements at different speeds and so it was necessary to normalize each plot to enable comparison of the kinematic movement patterns of two or more individuals13. The normalization forced the maxima of the primary movement values, to either side of the starting position, to occur at 25 and 75% of the time period, respectively, and the point of sign change between these two to occur at the 50% point. Linear interpolation between the original data points was used to give new values for the 100 data points of the measurement period. For each primary movement the accompanying movements in the other two planes were normalized according to the interpolation used for the primary movement.

25 26 23 24 26

Statistical analysis of the movements was then made possible. Means and standard deviations for the kinematic patterns of subject groups could be produced and differences between groups could be evaluated using a student t-test on each set of data points throughout the whole of the movement. Although comparisons were made throughout the movements the most significant differences occurred at the extreme positions as the patterns of movement were similar for all the groups.

Results The results are presented in three stages; patterns of kinematic movements, coupled movements and ranges of maximum movements.

Kinematic results

Plots of the mean movements, shown with + two standard deviations, of the youngest and oldest normal subject groups and all 80 normals together are shown in Figures 3-7. The movements in degrees are plotted against normalized time. (Flexion, left bend and left rotation are displayed as positive values; extension, right bend and right rotation as negative values.) These plots are reproduced with the resolution obtained from the graphics facilities on the personal computer used for this study. This was deemed appropriate as these are the plots to which patient data were compared (see below). The general patterns of movement were remarkably similar in all the groups. During both flexion and extension no appreciable lateral bend or axial rotation was seen, the mean values of these two movements remaining close to zero. During lateral bend a significant degree of opposite axial rotation was seen to occur, flexion was also seen to accompany the lateral bend to both sides. During axial rotation lateral bend was seen to occur but there was no consistent pattern of accompanying flexion or extension. These coupled movements are discussed in more detail later.

222

C/in.Biomech. 1990; 5: No 4

Figure 3. Males aged 20-29 yr. Mean and +2 standard deviations for the movements of this normal group plotted against normalized time during: a, flexion and extension; b, side bending to left and right; c, twisting to left and right (x-axis division = 1 s; y-axis division = 10’).

Sex differences in kinematics Each corresponding movement, in each age group, of males and females were compared with each other. A summary of the significant differences in kinematic movement patterns between the sexes is shown in Table 2.

Figure 4. Females aged 20-29 years. Mean and +2 standard deviations for movements of this normal group plotted against normalized time during: a, flexion and extension; b, side bending to left and right; c, twisting to left and right (axes scale as for Figure 3).

The table is best understood

with the aid of an ex-

ample. Consider the 40-49 year age group performing lateral bend. Female mobility is seen to be significantly greater than that of males in two planes; left bend, the primary movement, and flexion, an associated movement.

Hind/e et al.: Kinematics

of human back

223

t;

‘5 I-

a-

+E

Figure 5. Males s 50 yr. Mean and +2 standarddeviationsfor movements of this normal group plotted against normalized time during: a, flexion and extension; b, side bending to left and right;c, twisting to left and right (axes scale as for Figure 3).

Figure 6. Females 2 50 yr and over. Mean and f2 standard deviations for the movements of this normal group plotted against normalized time during: a, flexion and extension; b, side bendingto left and right:c, twistingto left and right(axes scale as for Figure 3).

The most striking point to emerge from this table is the very significantly greater flexion seen in 20-29 year old males in comparison with their female counterparts.

ences in the kinematic patterns and a summary of significant differences is presented in Tables 3 and 4. The overall impression one gets from these two tables is of a clear decrease in mobility with increasing age. The only exceptions to this occur in the female group where the > 50 yr age group is seen to exceed the 20-29 year age group in flexion. Interestingly the younger age group

Age differences

in kinematics

A similar analysis was performed

to examine age differ-

224

C/in. Biomech. 1990; 5: No 4 Table 2. Significant differences, at 5% level, in kinematic patterns of movement between sexes Groups tested 20-29

Movement petfonned Flex-ext Bend Twist Bend Twist Flex-ext Bend

yrs

30-39 yrs 40-49 yrs G ‘F I-

a-

2

Bend

50yrs

Difference M>F M>F F>M F>M F>M M>F F>M F>M F>M

Movements Flexion Flexion R Bend R Twist R Bend Flexion L Bend Flexion R Bend

*P < 0.01 Table 3. Significant differences, at 5% level, in kinematic patterns of movement with age in males

Age group

Movement performed

Difference

Movement

Flex-W

Y>O

Extension

Bend

Y>O Y > 0’ Y>O

L+ R Bend L Twist R Twist

Bend

Y>O Y>O

R+L Bend R+ L Twist

Twist

Y>O

L Bend

20-29 and 30-39

Bend

Y>O

L+ R Twist

30-39 and 2 50

Bend

Y>O

L+R Bend

40-49 and 2 50

Bend

Y>O Y>O

R Bend L Twist

20-29 and 2 50

Twist 20-29 and 40-49

Y = younger group 0 = older group *P < 0.01

Figure7. All 80 subjects combined. Mean and h2 standard deviation for movements plotted against normalized time during: a, flexion and extension; b, sida bending to left and right: c, twisting to left and right (axes scale as for Figure 3). demonstrates significantly greater extension so that the total range of flexion and extension was approximately the same in both groups and this suggests the lumbar lordosis increases with age in females. The older age group also shows increased twisting. Coupled movements

Subjectively a strong coupling of opposite axial rotation

upon lateral bend and vice versa can be seen by examining Figures 3-7. No out of plane movements are seen accompanying flexion and extension. Chi-squared tests confirmed a very strong coupling of opposite axial rotation on lateral bend and of opposite lateral bend on axial rotation. A strong coupling of flexion occurring with lateral bend was also shown, confirming the subjective impression one gets when viewing the plots of subjects performing lateral bend. No significant coupling, however, was seen between axial rotation and any sagittal plane movement. Regression analysis was then performed to establish the strength of this coupling. The magnitudes of the coefficients for all age groups were in the range O-3580.617 reflecting some correlation between the magnitude of the primary movement and the magnitude of the coupled movement. However, there was no con-

tiindle Table 4. Significant differences, patterns of movement with age Movement performed

at the 5% level, in kinematic

in females Difference

Movement

Flex-Ext

Y>O O>Y

Extension Flexion

Twist

O>Y Y>O

L Twist R Bend

Flex-Ext

Y>O

Extension

Bend

Y>O Y>O

R+L Bend R+L Twist

Twist

Y>O

R Bend

20-29 and 30-39

Twist

Y>O Y>O

L Twist R Bend

30-39 and 2 50

Bend

Y>O Y>O

L Bend R Twist

30-39 and 40-49

Twist

O>Y

R+L Twist

40-49 and 5 50

Bend

Y>O Y>O

R Bend LTwist

Age group 20-29 and 2 50

20-29 and 40-49

Table 5. Means of ranges of maximum voluntary

movements

Age group

Flexion Cl

Extension (“I

Lateral bend (“)

Axial rotation (“)

Male

20-29 30-39 40-49 3 50

74.6 73.2 77.2 70.1

26.0 16.7 23.5 19.4

57.9 53.0 47.4 37.5

30.3 30.0 29.1 21 .l

Female

20-29 30-39 40-49 2 50

59-4 70.3 64.0 73.0

31.6 24.0 19.8 21.1

61.9 53.6 52.8 50.5

31.8 25.8 36.6 29.3

Sex

sistent trend with age and there were no differences between the sexes. For coupling of flexion on lateral bending, there was little relation between the magnitude of the primary movement, lateral bend, and the magnitude of the coupled flexion. Again no trends were seen with respect to age and sex differences. Despite the uncertainty of the effects of age upon the strength of coupling, overall the coupled movements tended to be affected in the same manner as the primary movements, being reduced with age. Movement ranges

The means of the ranges of maximum voluntary flexion, extension, lateral bend and axial rotation for each of the

et al.: Kinematics

of human back

225

eight sample groups are displayed in Table 5. The values have been adjusted according to the regression equation presented previously and therefore represent true angles. When performing lateral bend and axial rotation there was large scale individual variation concerning the magnitude to which each movement was performed to the left and right. However, analysis (paired t-test) showed no consistent difference between left and right bend and left and right twist. Hence in Table 5, although sagittal plane movement is divided into flexion and extension, lateral bend and axial rotation are presented as the sum of the movements to the left and right.

Sex differences The means of the ranges of movement of all males and females are displayed in Table 6. This clearly shows males to have greater flexion but females to display more extension, lateral bend and axial rotation than their male counterparts. Analysis of variance (ANOVA) was employed to test these differences for significance. It showed that only in flexion was there any significant difference between males and females, with males having significantly more flexion than females (P < O-025). Age differences The effect of age upon ranges of maximum voluntary movement is illustrated in Table 5. It shows a general trend for decreasing movement with advancing age in all movements except flexion in the female groups, where there does appear to be a trend for increasing flexion with age. Only in lateral bend, for both sexes, is a consistent reduction in motion seen in each decade age group. The ANOVA analysis used above was also able to test for significant differences in the ranges of movement between age groups. This analysis confirmed that the trend displayed in lateral bend is, indeed, significant (P < 0.025). It also showed extension to be significantly reduced with age (P < 0.05). No significant differences were observed in flexion or twist. Regression analysis was then carried out to determine the relationship between the age and the range of movement of a subject. There was a poor relationship between age and flexion, especially in the female group where the correlation coefficient of r = 0.339, indicated an increase in flexion with advancing age. For all other movements, apart from axial rotation in females, the negative coefficients

Table 6. Means of ranges of movement females

Sex

Male Female

for all males and

Flexion PI

Extension (“)

Lateral bend

Axial rotation

(“I

(“I

73.8 66.7

21.5 24.1

48.4 54.1

27-l 30.8

226

C/in. Biomech. 1990; 5: No 4

indicated a reduction of movement with age. The strongest relationship was seen to occur with lateral bend where the correlation coefficient for all subjects combined was r = -0.425. Between the sexes there were no consistent differences in the reduction of motion with age apart from flexion which was reduced in males but increased in females, although neither trend reached statistical significance. The ANOVA employed earlier indicated no significant sex-age combination effect for any of the movements. Clinical study Following the successful study of normal kinematic movement patterns of the human back, a preliminary clinical study was undertaken in order to assess the ~SPACE ISOTRAK in a clinical setting. A detailed description of this study is given elsewhere*s. Thirty-six patients who were being considered for back surgery were measured in two orthopaedic outpatient clinics. The basic measurement procedure remained the same as that described in the collection of normal data. In addition, the two surgeons involved completed assessment forms for each patient including a subjective clinical assessment of mobility. There was a large range of pathological conditions present in the patients but combining them together in general groupings of those with clinically similar features permitted some observations to be made. Four groups with at least five individuals in each were identified from the two clinics. There were two groups in whom the prime source of pain was diagnosed as being associated with an intervertebral disc, one group in whom the prime source was associated with lateral recess restriction and one group in whom zygapophysial joint arthropathy was diagnosed.

Results of clinical study Comparison of the movement patterns of each patient to a normal group matched for age and sex showed that all the patients had widespread and marked deviations from normal in both primary and coupled movements. The four groupings all included both male and female patients in several age ranges and so the mean movements of each group were compared to those of all 80 normals together (Figure 7). The groups tested showed identifiable and different deviations from normal patterns of movement. A summary of these differences is presented in Table 7 which shows that most of the differences occurred in lateral bending and axial rotation and their associated out-ofplane movements. These differences were not just in the magnitudes of movement but in the relationship between the primary attempted movement and the movements in the other two planes. Although significant differences from the normal patterns were found, the differences between the groups did not reach statistical significance.

Table 7. Summary of significant differences from normal, at 5% level, of the kinematic movement patterns for the patients Lateral bend Patient group Disc 1 Disc 2 Lateral recess Zygapophysial joint

Axial rotation

Primary Assoc Primary Assoc Flex fxt /at bend ax rot ax rot /at bend 0 + + +

+ + + +

z, + +

0 0 + +

0 + A

+ A 0

+ = Restricted 0 = Normal

A further important observation was that therz was only poor correlation between the site of pain and abnormalities in lateral bending and axial rotation to that side. The clinicians’ observations of movements drd not correlate well with the measurements. In particular, several patients who had significant restriction of some movements were recorded by the clinicians as having normal mobility. Discussion The term ‘normal’ is somewhat inappropriate when discussing backs. If everyone who had experienced any back trouble was excluded from the normal study there would have been great difficulty in completing this sample, especially in the older age groups. Bearing this in mind the criteria for acceptable subjects were that they should have been free from low back pain for the previous six months and should not have undergone surgery at any time. The results indicate that the ~SPACEISOTRAKmeasured ranges of motion in excess of those known to occur in the lumbar spine of healthy individuals”. The exaggeration of true lumbar movement resulting from the attachment of markers to the skin was discussed earlier. Inevitably, with the necessary tethering of the sensor with the strap around the trunk, the ISOTRAK can only ever claim to give a measure of ‘low back’ mobility which must include some thoracic movement. Further attention to the method of sensor attachment may be appropriate, however, there is no reason to believe that these measurements do not give a fair representation of the actual movement of the spine. Indeed, the patterns and coupling of movements observed in this study compared to those that have been shown to occur in the lumbar spine” would tend to support this argument. There is little information available concerning the normal kinematic patterns of movement of the lumbar spine. The VICON system (an opto-electronic device) was used to look at patterns of movement in six normal individuals’*. All subjects displayed consistent patterns of movement and there was a strong similarity of these patterns to those obtained from the three-dimensional radiographic studies ” , leading to the conclusion that

Hind/e et al.: Kinematics

surface measurements are closely related to the movements of the underlying spine. Typical plots obtained for flexion-extension, lateral bend and axial rotation from the subjects in this study show very good agreement with those obtained from the three-dimensional radiographic study and the VICON study. A number of authors have previously reported age and sex differences with respect to ranges of normal lum. Decrease in mobility with age has also bar mobility 29-33

been reported by researchers conducting in vitro studies of lumbar spine motion34’3’. The results of this study agree broadly with the consensus evidence of previous studies, A general trend was observed for decreasing mobility with increasing age, the only exception to the trend was the significant increase in sagittal flexion observed with age in the female group. This result would seem to be a consequence of an increase in lordosis with increasing age in the female group, as was noted in the results. The observation of coupled lateral bend with axial rotation and vice versa in this study agrees broadly with the coupling Pearcy noted at the intervertebral level”. He also found no significant coupling of flexion or extension with axial rotation. However, he noted consistent coupled extension on lateral bend at all levels bar LsSl . This study showed a consistent coupling of flexion with lateral bend. Pearcy et al.‘* again found this trend for extension coupled with lateral bend during their kinematic studies of back movement using the VICON system, although they did note llexion occurring in three of their subjects, half the sample size, in at least one of the right and left bends. The reason for this discrepancy lies in the experimental technique employed for both the threedimensional radiographic and the VICON studies. In both cases the subjects were required to stand in frames which held the subjects’ anterior superior iliac spines against moulded plastic pads. So positioned, a subject’s movements were artificially restrained preventing the natural coupling of flexion with lateral bend demonstrated in this study, where they were able to move freely. The VICON study did show opposite axial rotation with lateral bend and vice versa. It is reasonable to assume, therefore, that the coupling characteristics demonstrated in this study are a true reflection of those occurring in the normal unrestrained human back. The patient study was conducted as a preliminary trial to assess the use of the ISOTRAK in a hospital clinic. It proved easily adaptable for use in the clinical setting being quick to set up and requiring relatively little space in which to operate. Furthermore, individual patient contact time was reasonably short at approximately 10 min. The movements of the patients demonstrated that individuals with back problems do have marked abnormalities in their movements both in the primary movement and the pattern of accompanying, out-of-plane, movements. These results suggest that the ISOTRAK is able to distinguish between the kinematic patterns ofthe normal and pathological spine, for even in the small heterogeneous groups investigated in this pilot study identifiable patterns of movement were seen. A further

of human back

227

study is now underway to assess groups of patients with similar symptoms and diagnoses. This will establish whether a pathology is identifiable by its effect on movements and the relation, if any, between segmental disease and overall back movement. The discrepancy between the clinicians’ observations of movement and the measurements highlights the problem of identifying three-dimensional movement by eye and suggests that the use of such observations be reassessed. In conclusion these investigations have shown that the ~SPACE ISOTRAK is a device capable of accurate and reproducible measurements of three-dimensional movements. A comprehensive assessment of normal low back mobility, has been produced, showing normal subjects to have clearly identified patterns of movement. Indeed, the similarity of normal kinematic patterns observed across the range of age and sex groups in this study was remarkable. Low back pain patients showed widespread disruption to their primary and coupled movement patterns and when grouped together by diagnosis showed evidence of discrete and identifiable alterations from normal. However, further studies of homogeneous patient groups are required in order to demonstrate whether these measurements are of actual clinical use. Finally, this study has shown that the ISOTRAK fulfills the criteria for a clinical measurement device. It is an effective and relatively inexpensive method for the non-invasive three-dimensional kinematic measurement of low back mobility. Acknowledgements

We wish to thank all part in these studies, patient data collation Research Council for

volunteers and patients who took Mr F H Khan for assisting with and the Science and Engineering financing part of the work.

References 1 Pearcy MJ. Measurement of back and spinal mobility. Clin Biomech 1986; 1: 44-51 2 Portek I, Pearcy MJ, Reader GP, Mowat AG. Correlation between radiographic and clinical measurement of lumbar spine movement. Br J Rheumatol 1983; 22: 197-205 3 Reynolds PMG. Measurement of spinal mobility: a comparison of three methods. Rheumatol Rehabil 1975; 12: 180-5 4 Moll J, Wright V. Measurement of spinal movement. In: Jayson M, ed. The Lumbar Spine and Back Pain. London: Sector Publishing Ltd., 1976: 93-112 5 Mellin GP. Comparison between tape measurements of forward and lateral flexion of the spine. Clin Biomech 1989; 4: 121-3 6 Pearcy MJ, Portek I, Sheperd JE. The effect of low back pain on lumbar spine movements by three-dimensional x-ray analysis. Spine 1985; 10: 150-3 7 Stokes IAF, Wilder DG, Frymoyer JW, Pope MJ. Assessment of patients with low back pain by biplanar radiographic measurements of intervertebral motion. Spine 1980; 6: 233-40 8 Auchincloss S. The Painful Back, Practical Aspects of Management. Oxford: Medicine Publishing Foundation, 1983

228

C/in. Biomech. 1990; 5: No 4

9 Roland M. Back pain . . . an unsolved problem. Talkback

10 11 12

13

14

15

16

17

18

19

20

21

1983; 14: 180 Wells N. Back Pain. London: Publication No. 78. London Office of Health Economics, 1985 Pearcy MJ. Stereo radiography of lumbar spine motion. Acta Orthop Stand 1985; 56, Suppl212 Pearcy MJ, Gill JM, Hindle RJ, Johnson GR. Measurement of human back movements in three dimensions by opto-electronic devices. Clin Biomech 1987; 2: 199-204 Pearcy MJ, Hindle RJ. New method for the non-invasive three-dimensional measurement of human back movement. Clin Biomech 1989; 4: 73-9 Krieg JC. Electromagnetic head tracking, a new communication device for persons who are severely disabled. Proc Interagency Conf Rehabil Eng, Atlanta, Georgia, 1979; 26-31 Krieg JC. A feedback mechanism for use in paraplegic stimulation techniques. Proc 2nd Int Conf Rehabil Eng, Ottawa, 1984 An KNJ, Jacobsen MC, Berglund LJ, Chao EYS. Application of a magnetic tracking device to kinesiologic studies. J Biomech 1988; 21: 613-20 Buchaiter DN, Pamianpour M, Nordin M, Kahnovitz N. A non-invasive in-vivo technique for examining posture and functional spinal motion. In: Karwowski W, ed. Trends in Ergonomic&h.tman Factors III. Elsevier, 1986: 727-37 Elnagger M, Nordin M, Sheikhzadel A, Kahnovitz N. Correlation between low back pain severity and spinal mobility in chronic low back pain patients treated with spinal exercises. Presented at International Society for the Study of the Lumbar Spine, Miami, Florida, April 13-17, 1988 Elnagger M, Nordin M, Sheikhzadel A, Kahnovitz N. The effects of spinal flexion exercises on low back pain and spinal mobility in chronic mechanical low back pain patients. Presented at the International Society for the Study of the Lumbar Spine, Miami, Florida, April 13-17, 1988 Buchalter D, Pamianpour M, Viola K, Nordin M, Kahanovitz N. Three-dimensional spinal motion measurements. Part 1. A technique for examining posture and functional spinal motion. J Spin Disord 1989; 1: 279-83 Buchalter D, Kahanovitz N, Viola K, Dorsky S, Nordin M. Three-dimensional spinal motion measurements. Part 2. A

22 23

24

25

26

27

28

29

30

31 32 33

34 35

non-invasive assessment of lumbar brace immobilisation of the spine. J Spin Disord 1989; 1: 284-6 Burton AK. Regional lumbar sagittal mobility; measurements by flexicurves. Clin Biomech 1986; 1: 20-6 Burton AK. Patterns of lumbar sagittal mobility and their predictive value in the natural history of back and sciatic pain [PhD thesis]. Huddersfield Polytechnic, UK, 1987 McConnell DG, Beal MC, Dirmar U, Goodridge JP, Johnston WL, Komi Z, Upledger JE, Blum G. Low agreement of findings in neuromuscular skeletal examination by a group of osteopathic physicians using their own procedures. J Am Osteopath Ass 1980; 79: 441-50 MacGibbon B, Farfan HF. A radiographic survey of various configurations of the lumbar spine. Spine 1979; 4: 258-66 Stokes IAF. Relationship between movements of vertebrae and adjacent skin markers and sections of the back. In: Annual Report, Oxford Orthopaedic Engineering Centre, Headington, Oxford, UK. 1977: 41-2 Pearcy MJ, Gill JM, Whittle MW, Johnson GR. Dynamic back movement measured using a three-dimensional television system. J Biomech 1987; 20: 943-9 Hindle RJ. Three-dimensional kinematics of the normal and pathologic spine [PhD thesis]. University of Durham, UK, 1989 Batti’e MC, Bigos SJ, Sheehy A, Wortley MD. Spinal flexibility and individual factors that influnence it. Phys Ther 1987; 67: 653-8 Fitzgerald GK, Wynkeen KJ, Rheault W, Rothschild B. Objective assessment with establishment of normal values of lumbar spinal range of motion. Phys Ther 1983; 63: 1776-81 Mall JMH, Wright V. Normal range of spinal mobility. Ann Rheum Dis 1971; 30: 381-6 Tanz SS. Motion of the lumbar spine. A roentgenologic study. Ann J Roentgen 1953; 69: 399-412 Troup JDG, Hood CA, Chapman AE. Measurements of sagittal mobility of the lumbar spine and hips. Ann Phys Med 1968; 9: 308-21 Hilton RC, Ball J, Ben RT. In vitro mobility of the lumbar spine. Ann Rheum Dis 1979; 38: 378-83 . . Taylor J, Twomey J. Sagittal and horizontal plane movement of the human lumbar vertebral column in cadavers and in living. Rheumatol Rehab 1980; 19: 223-32