AIM

When a hip is replaced using a posterior surgical approach, some of the external rotator muscles are divided. The aim of this study was to assess if this surgery has a long term affect on hip rotation during activities of daily living.

Introduction: The aim of this study was to compare hip movement between normal subjects and patients with a large Metal on Metal hip replacement, undertaking the task of retrieving an object from the floor.

Methods: An electromagnetic tracker was used to measure movement as subjects retrieved an object with flexed hips and straight knees. Measurements were taken from a control group of 10 subjects with bilaterally normal hips, and 10 subjects with unilateral hip replacement. Sensors were attached over the iliac crest and the mid-shaft of the lateral thigh. Data was collected as each subject repeated the movement 3 times. The tracker recorded data at 10 hertz, with an accuracy of 0.15 degree.

Results: For the normal group the mean hip flexion was 90.8 degrees (SD 20.1). For the arthroplasty group the mean flexion on the normal and operated sides were 74.0 (SD 21) and 72.7 degrees (SD 21) respectively. This was not significant (P= 0.83). However there was a significant difference in hip movement between the operated hips and those in the normal control group (P= 0.03).

For the bilaterally normal group the mean hip rotation was 2.9 degrees internal (SD 11.8). For the arthroplasty group the mean rotation on the normal and operated sides were 9.4 degrees external (SD 9.5) and 6.9 degrees internal (SD 13.9) respectively. In this group there was a significant difference between the normal and operated side (P= 0.02).

Discussion: This study has shown that patients with a unilateral hip replacement have no significant flexion difference between hips, when retrieving an object from the floor. However there was a significant difference compared to a control group with normal hips. A significant difference was also observed when comparing the rotation of an operated hip joint to the contra-lateral normal hip in the same individual.

Introduction: Patients undergoing total hip arthroplasty are advised to minimise their hip flexion in the early postoperative phase, to reduce the risk of dislocation. One activity that requires hip flexion is picking an object up from the floor. The aim of this study was to investigate the amount of hip flexion required to perform this task, and to see if there is a difference between patients with small and large bearing total hip replacements.

Methods: Nineteen unilateral total hip replacement patients were recruited into the study. Nine had a small bearing (metal on plastic) implant and ten had a large bearing (metal on metal) implant. Each patient had a contra-lateral normal native hip, which provided a control for bilateral comparison.

An electromagnetic tracking system was used to measure the flexion in the operated and normal hip of each patient. Tracker sensors were placed on the iliac crest and the mid-lateral thigh. The patients were then asked to flex forward from a standing position to pick an object up off the floor. This movement was repeated 3 times. Flexion data was collected at 10Hz which was accurate to 0.15 degrees. Spinal flexion was not recorded during the task.

Patients were also asked to complete the Harris and Oxford Hip Score questionnaires to obtain qualitative data regarding their hip replacement.

Results: The mean peak flexion angles (degrees) for each group are given below:

Small bearing group:

Operated side: Peak flexion = 79.3

Comparing the small bearing group with the large bearing group, the peak difference was 6.6. This difference was non-significant with P = 0.43.

All patients reported good – excellent functional results when completing the Harris and Oxford Hip Scores.

Discussion: The investigation showed that picking an object up from the floor requires a peak hip flexion of approximately 80 degrees. This investigation found no significant difference between the normal and operated sides. This would suggest that a Total Hip Replacement restores the “normal” range of motion in a hip joint. Furthermore, there was no significant difference between the small and large bearing hip implants.

Introduction: Hip replacement patients are prone to dislocations during extreme hip movement in the early post operative period. An activity of daily living that puts them at risk of dislocation is picking an object off the floor. The aim of this study was to assess the movement of the hip using different techniques to pick an object of the floor.

Methods: An electromagnetic tracking system was used to assess hip movements for four different techniques in picking an object from the floor. These were -

Flexing forward to pick an object up between the feet.

Measurements were taken from 40 hips in 20 normal subjects aged 21 to 61. Sensors were attached over the iliac crest and the mid-shaft of the lateral thigh. Data was then collected from the magnetic tracker as each technique was repeated 3 times. The system recorded hip flexion and rotation data at 10 hertz, with an accuracy better than 1 degree. Data was then analysed and the mean readings for each technique were compared.

Results: For each of the four techniques listed above the respective mean (SD) results were:

Flexion: 81.4 (27.5), 83.3 (27.6), 93.3 (28.7) and 33.5 (17.6) degrees.

The most significant movements for each technique were flexion and external rotation.

The movements with the least and most flexion were kneeling (33.5 deg) and squatting (93.3 deg). They were significantly different with a paired t-test p< < 0.001.

The movement with the least and most external rotation were kneeling (7.5 deg) and side pick up (22.7 deg). They were significantly different with a paired t-test, p< < 0.001.

Conclusion: This study has found that the most effective technique to pick up an object from the floor is kneeling as this has the least amount of flexion and external rotation. We conclude that this is the safest technique in carrying out this activity in the early post operative stage for patients who have undergone a total hip replacement.

Introduction: Leg length discrepancy (LLD) following hip arthroplasty can produce abnormal loading leading to pain, increased wear and loosening of implants. The aim of this study was to investigate the relationship between LLD and static limb loading.

Methods: A pedobarograph was used to measure the limb loading of 19 normal volunteers aged 18 to 58. Each volunteer was asked to stand on the Pedobarograph with both feet so that their weight could be recorded. The load through the left leg was then recorded with the right leg on a platform level beside it. The platform was then raised in 1 cm increments to 6 cm, to simulate different levels of LLD. In each position 3 readings were taken with the right knee flexed (pelvis level), and straight (pelvis tilted).

Results: When the feet were level the left leg took 53 % of the load. As the height of the right foot was increased the load through the left leg increased in a non-linear fashion.

With the knee flexed, a 1 cm difference produced a 3 % increase in loading. This was significant (P< 0.05). All subsequent increases were also significant. The largest increase in load was observed between 1 cm & 2 cm (+5 %). At 6cm the left leg load was 70.9 %.

With the pelvis tilted, there were smaller increases in loading. These did not become significant until a difference of 5 cm. The maximum load was 62.1 % at 6 cm.

Discussion: The length-loading relationship was non-linear. The pelvis tilted stance produced less loading asymmetry, but more discomfort than the flexed knee stance.

Introduction: Activities that require extreme hip movement can dislocate hip implants in the early post operative phase. The aim of this study was to assess the movement of the hip using four different techniques to retrieve an object from the floor.

Methods: An electromagnetic tracker was used to measure hip movement during these retrieval techniques:-

Flexing forward to pick up an object between the feet

Measurements were taken from 50 hips in 25 normal subjects aged 21 to 61. Sensors were attached over the iliac crest and the mid-shaft of the lateral thigh. Data was collected as each technique was repeated 3 times. The tracker recorded hip flexion and rotation data at 10 hertz, with an accuracy of 0.15 degree.

Results: For each of the four techniques the respective mean (SD) movements were:-

Flexion: 75.8(28.6), 79.2(27.2), 87.5(29.7) and 30.4(17.3).

Kneeling had significantly less flexion and external rotation than all the other techniques (paired t-test, P< < 0.001).

Discussion: Flexion and external rotation were the most significant movements for each technique. The movements with the least and most flexion were kneeling (30.40) and squatting (87.50). The movement with the least and most external rotation were kneeling (7.30) and side pick up (20.10).

Kneeling has the least amount of movement, therefore, it minimises the risk of dislocation when retrieving an object from the floor.

Introduction: Posterior dislocation of replacement hips may occur during extreme hip flexion and adduction. Hip braces restrict movement, but they are uncomfortable and have a low patient compliance. Knee braces are more comfortable, and also restrict hip movement, by tightening the hamstrings. This study investigated the effect of a knee brace on hip movement.

Methods: A magnetic tracker was used to measure the movement of 20 normal hips in 20 volunteers, aged 25–62. Sensors were attached over the iliac spine and lateral thigh. Subjects were asked to lie on a couch and flex and adduct their hip three times with their knee bent and three times with their knee braced in extension. During each movement the tracker recorded hip flexion and adduction angles, with an accuracy of 0.15 degrees.

Results: With the knee flexed, the mean hip flexion angle was 66.00 (SD 11.0). With the knee braced, the mean hip flexion angle was 35.30 (SD 15.4). Hence the knee brace reduced hip flexion by 46 % (30.70) (paired t-test, P < < 0.001).

With the knee flexed, the mean hip adduction angle was 23.70 (SD 7.1). With the knee braced, the mean hip adduction angle was 21.60 (SD 5.6). Hence the knee brace reduced hip adduction by 9 % (2.10). This was not significant (paired t-test, P = 0.3).

Discussion: These results indicate that a knee brace can restrict hip flexion by almost 50%. This information may be useful for patients in whom restriction of hip flexion provides hip stability. As the knee brace is more comfortable than the hip brace, a better patient compliance is expected.

Aim: The carpal bone arrangement can be described as a matrix of two rows and three columns. There a various theories as to how the bones within the matrix move during ulna to radial deviation. One theory suggests that there are two types of wrist movement, namely Row & Column¹.

The aim of this study was to investigation how the rotational axis of the wrist moves as the hand goes from full ulna to full radial deviation.

Materials and Methods: Ulna to radial deviation was assessed in 50 normal wrists in 25 normal subjects aged 19 to 57. Movement was measured with a Polhemus Fastrak (TM) magnetic tracking system. The system has translational and rotational measurement accuracies of 1 mm and 1 degree respectively. Subjects placed their palms on a flat wooded stool and had movement sensors attached over their 3rd metcarpal and distal radius. These sensors then recorded movement as the hand moved from full ulna to full radial deviation.

Results: The mean range of movement was 45 degrees (SD 7). In full ulna deviation the wrist rotational axis was in the region of the lunate. As the hand moved towards radial deviation, the axis moved distally. At the end of the movement the mean distal displacement was 21 mm (SD 15). In 32 wrists the distal displacement was accompanied by mean displacement towards the ulna of 12 mm (SD 8). In 18 wrists the distal displacement was accompanied by a mean displacement towards the radius of 8 mm (SD 5).

Conclusion: The rotational axis position indicates how the wrist is moving during radial deviation. In early movement, when the axis is proximal, there is a high degree of sideways translation. In later movement, when the axis is distal, there is more rotational movement. In some cases the axis moved distally and toward the radius, whereas in other cases it moved distally and toward the ulna. This spectrum of movement may support the theory of 2 types of carpal movement. i.e. Column movers and row movers¹.

Introduction: Posterior dislocation of replacement hip joints may occur during hip flexion and adduction. A hip brace is commonly used for recurrent dislocations in patients awaiting revision surgery or when unfit for it. However, these hip braces are cumbersome and have a low patient compliance.

Knee braces are more comfortable to wear, and they also restrict hip movement by tightening the hamstrings. With this background we investigated the effect of a knee brace, applied in full extension, on hip flexion and adduction.

Methods: The movement of 20 normal hips in 20 healthy volunteers aged 25–62, were assessed using a magnetic tracking system (Polhemus Fastrak). One tracking sensor was attached near the anterior superior iliac spine and another one on the lateral aspect of the thigh at a fixed distance from the knee joint. Subjects were then asked to lie on a couch and flex and adduct their hip three times each with the knee bent and then with their knee braced in extension. Two sets of three readings were recorded. During each movement the tracker recorded hip flexion and adduction angles, with a measurement accuracy of 0.15 degrees.

Results: With a flexed knee, the mean hip flexion angle was 66.0 degrees (CI95 = 61.1, 70.8). With the knee braced, the mean hip flexion angle was 35.3 (CI95 = 28.5, 42.1). Hence the knee brace reduced hip flexion by 46% (30.7 deg). A paired t-test found this highly significant, with P < < 0.001.

With a flexed knee, the mean hip adduction angle was 23.7 degrees (CI95 = 20.6, 26.9). With the knee braced, the mean hip adduction angle was 21.6 (CI95 = 19.2, 24.1). Hence the knee brace reduced hip adduction by 9% (2.1 deg). A paired t-test found this was not significant with P = 0.3.

Conclusion: The results indicate that a knee brace can restrict hip flexion by almost 50%. This information may be useful for patients in whom restriction of hip flexion provides hip stability. As the knee brace is more comfortable than the hip brace, a better patient compliance is expected.

Leg length discrepancy (LLD) is a recognised complication of total hip arthroplasty. LLDs can cause abnormal weight bearing, leading to increased wear, aseptic loosening of replacement hips and pain. To compensate for LLDs the patient can either flex the knee of the long leg or tilt their pelvis. The aim of this project was to investigate how stance affects static limb loading of patients with leg length discrepancy.

A pedobarograph was used to measure the limb loading of 20 normal volunteers aged 19 to 60. A 2 second recording with both feet on was taken to establish their body weight. Readings were taken of the left foot with the right level, 3.5cm lower (simulating a long left leg) and 3.5cm higher. In each case three readings were taken with the knee flexed and three readings with the knee extended.

When both feet were at the same level, the left limb took 54% of the load.

When the right foot was lower and the left knee flexed, the left leg took 39 % of the load (P < 0.001) (paired t-test). When the left knee was extended the left leg took 49 % of the load (P = 0.074).

With the right foot higher and right knee flexed, the left leg took 65 % of the load (P < 0.001). When the right knee was extended the left leg took 58 % of the load (P = 0.069).

These results show that weight distribution is increased in the simulated shorter limb. Loading is greater when the longer limb is flexed. Tilting the pelvis reduced the load. However this may cause pelvic and spinal problems.

Uneven load distribution is likely to lead to early fatigue when standing and may explain why some post arthroplasty patients with limb length discrepancy have poor outcomes.

Posterior dislocation of replacement hips may occur during hip flexion and adduction. Whilst hip braces can restrict hip movement, they are cumbersome and have a low patient compliance. Knee braces are more comfortable to wear and also restrict hip movement by tightening the hamstrings. This study investigated the effect of a knee brace on hip flexion and adduction.

The movement of 20 normal hips in 20 healthy volunteers aged 25–62, were assessed using a magnetic tracking system (Polhemus Fastrak). Tracking sensors were attached over the iliac crest and lateral thigh. Subjects were asked to lie on a couch and flex and adduct their hip three times with their knee bent. A knee brace was then applied and the hip movements were repeated with the knee extended. During each movement the tracker recorded hip flexion and adduction angles with an accuracy of 0.15 degrees.

When the knee was flexed, the mean hip flexion angle was 66.00 (CI95 = 61.1, 70.8). When the knee was braced, the mean hip flexion angle was 35.30 (CI95 = 28.5, 42.1). Hence the knee brace reduced hip flexion by 46 % (30.70). A paired t-test found this highly significant (P < 0.001).

When the knee was flexed, the mean hip adduction angle was 23.70 (CI95 = 20.6, 26.9). When the knee was braced, the mean hip adduction angle was 21.60 (CI95 = 19.2, 24.1). Hence the knee brace reduced hip adduction by 9 % (2.10). A paired t-test found this was not significant (P = 0.3).

These results indicate that a knee brace can restrict hip flexion by almost 50%. This information may be useful for patients in whom restriction of hip flexion provides hip stability. As the knee brace is more comfortable than the hip brace, a better patient compliance can be expected.

Aim: The aim of this investigation was to determine how the rotational axis of the wrist moves as the hand goes from full ulna to full radial deviation.

Materials & Methods: Ulna to radial deviation was assessed in 30 normal wrists in 15 normal subjects aged 19 to 32. Movement was measured with a Polhemus Fastrak (TM) magnetic tracking system. The system has translational and rotational measurement accuracies of 1 mm and 1 degree respectively. Subjects placed their palms on a flat wooded stool and had movement sensors attached over their 3rd metacarpal and distal radius. These sensors then recorded movement as the hand moved from full ulna to full radial deviation.

Results: The mean range of movement was 47 degrees (SD 8). In full ulna deviation the wrist rotational axis was in the region of the lunate/capitate. As the hand moved towards radial deviation, the axis moved distally. At the end of the movement the mean distal displacement was 22 mm (SD 14). In 17 wrists the distal displacement was accompanied by mean displacement towards the ulna of 13 mm (SD 8). In 13 wrists the distal displacement was accompanied by a mean displacement towards the radius of 7 mm (SD 5).

Conclusion: The rotational axis position indicates how the wrist is moving during radial deviation. In early movement, when the axis is proximal, there is a high degree of sideways translation. In later movement, when the axis is distal, there is more rotational movement. In some cases the axis moved distally and toward the radius, whereas in other cases it moved distally and toward the ulna. This spectrum of movement may support the theory of 2 type of carpal movement. i.e. Column movers and row movers [Craigen & Stanley].

Aim: One of the issues of metatarsophalangeal joint (MTPJ) replacements is that they do not restore full range of movement (RO M). However, full RO M is not needed for functional walking. The aim of this study was to measure the difference between the functional and maximum ROM of the first metatarsophalangeal joint.

Materials & Method: The functional and maximum ROM of 32 MTPJs in 16 normal adults were measured with a video imaging system. The system first measured the ROM as the subject walked past the camera. It then measured the ROM as standing maximum extension tests were performed.

Results: During functional walking tests the mean ROM was 37.9 degrees (SD 12.2). During maximum standing extension tests the mean ROM was 64.9 degrees (SD 11.3). Therefore the functional walking ROM was only 58% of the maximum standing extension ROM, with a mean difference of 27 degrees. A paired t-test comparison showed P< 0.0001.

Discussion: MTPJ arthroplasty has previously been criticised because it does not restore full RO M. However, the results of this study suggest that the functional movements required in normal gait are significantly less than what can be maximally achieved in clinical standing extension tests. Therefore arthroplasty can be a suitable treatment if it can provide an adequate functional RO M.

Conclusion: The results of this study show that the functional range of movement required for walking is only 58% of the maximum extension ROM of the first MTP joint. Therefore, MTP joint replacements do not need to restore maximum extension, as normal gait can be achieved without this.

Introduction: Sock application is a daily task that can pose a dislocation risk to implanted hips. The aim of this study was to measure hip flexion and rotation using three seated techniques of sock application. Namely:- 1. The leg crossed in a “figure of four” technique, 2. The lean forward technique, and 3. using a specialised sock applicator (Foxy Sock Aid).

Methods: The movement of 32 hips in 16 healthy male volunteers, aged 20–43, were assessed as socks were applied using the 3 techniques listed above. Hip flexion and axial rotation were measured with a “Polhemus Fastrak” magnetic tracking system. Data was recorded from magnetic sensors attached around the subjects femur and Iliac Spine. The sensors measurement accuracy was 0.15 degrees.

Results: All measurements started with the subjects sitting with their hips in approximately 90 degrees of flexion. The mean (SD) additional flexion required for each of the 3 techniques were:- Cross leg = 57.2⁰ (13.7), Lean forward = 51.3⁰ (17.7), and Sock applicator = 11.2⁰ (7.2). Hence the sock applicator required 40.1 degrees less flexion (P< < 0.001) than the lean forward technique. The lean forward technique required 5.9 degrees less flexion (P=0.007) than the cross legged technique.

The mean (SD) peak rotations for each of the 3 techniques were:- Cross leg = 35.1⁰ (9.8) external, Lean forward = 0.1⁰ (3.8) internal, and Sock applicator = 0.8⁰ (4.0) internal.

Discussion: The sock applicator, when used correctly, requires very little hip flexion. When patients stop using the sock applicator, they should be encourage to use the lean forward technique, as this requires little axial rotation.

Introduction: A common outcome measure for Total Knee Replacement patients is the measurement of their knee extension angle. In theatre the surgeon usually ensures that the replacement knee can achieve full extension with the patient supine. However patients frequently comment that they are unable to reach full extension while seated. This is due to the flexed hip tightening the hamstrings.

The aim of this study was to deduce the effect of Hip flexion on the knee’s full extension angle (popliteal angle) in a control group of subjects with normal knees.

Method: An electromagnetic motion analysis system (Polhemus Fastrak) was used to assess twenty knees in ten normal subjects. The groups mean age, height and weight were 35 years (SD=7), 1.82 m (SD=0.05) and 83.9 kg (SD=12.9) respectively. Each subject was placed in a supine position on an examination couch with their legs hanging over the end and their knees in 90 degrees of flexion. The electromagnetic source was then positioned 50 cm from the knee joint. Two electromagnetic sensors were then taped to the lateral side of the leg, one over the femoral midpoint and one over the tibial mid point. A recording was then made as the subject extended their knee to full extension. The subjects hip was then flexed to 90 degrees with the knee in 90 degrees flexion. The subject was the asked to extend their knee as far as possible, while keeping their hip flexed. This processed was repeated 3 times for each knee to give average knee extension angles, with the hip straight and flexed.

Results: In the supine hip-straight position the mean extension angle for the fully extended knees was 1.2° (SD=2.7°). In the supine hip-flexed position the mean extension angle for the fully extended knees was 23.8° (SD=12.6°). This gave a mean difference of 22.6°. A paired t test of the extension angles for the two hip positions yielded a significant difference with p=0.0000001.

Conclusion: These results indicate that hip flexion significantly reduces the amount of full knee extension, in normal subjects.

Introduction: The Lachman test for anterior cruciate ligament (ACL) deficiency, requires a subjective assessment of joint movement, as the tibia is pulled anteriorly. This study has objectively quantified this movement using a magnetic tracking device.

Materials and Methods: Ten patients aged 21 to 51 years were assessed as having unilateral ACL deficiency with conventional clinical tests. These patients were then reassessed using a magnetic tracking device (Polhemus Fastrak). Patients had magnetic sensors attached around the femoral and tibial mid-shafts using elasticated Velcro straps. The Lachman test was then performed with the patient lying within range of the system’s magnetic source. The test was performed three times on the normal and injured knees of each patient. During the tests, sensor position and orientation data was collected with an accuracy better than 1 mm and 1 degree, respectively. The data was sampled at 10Hz and stored on a computer for post-test analysis. This analysis deduced the tibial displacement resulting from each Lachman pull.

Results: The main Lachman movement is an anterior displacement of the tibia with respect to the femur. The mean anterior movement for the normal knees was 5.6 mm (SD=2.5). By comparison the ACL deficient knees had a mean anterior movement of 10.2 mm (SD=4.2). This is 82 % more. A paired t test of this data showed it to be highly significant with P = 0.005. In addition to the anterior movement, there was also a small proximal tibial movement. In the normal knees the mean movement was 0.7 mm (SD=1.9). In the injured knees the mean movement was 2.1 mm (SD=3.4). However, this difference was not significant (P = 0.12).

Conclusion: This study has quantified the movement produced during the Lachman test for ACL deficiency. The results compare well with reported results from similar arthrometer tests[1]. The main advantage of the magnetic tracker is that its lightweight sensors cause minimal disturbance to the established clinical test. It therefore offers a convenient and non-invasive method of investigation.

Introduction: Pedobarograph systems are used to measure foot pressure characteristics during gait. These measurements help clinicians diagnose pathology and assess treatment outcome. While most patients can walk across the Pedobarograph footplate unaided, some patients ask if they can use their crutch.

Aim: The aim of this study was to assess the effect of using a crutch on pedobarograh measurements.

Materials and Methods: A Musgrave Pedobarograph system was used to measure the foot pressure characteristics of twenty feet in ten normal subjects. The group’s mean age and weight were 31 years (SD=6) and 78 kg (SD=12), respectively. Each subject had each of their feet measured 3 times as they walked across the foot plate unaided. These measurements were then repeated with the subject using a crutch on the contralateral side to that being measured. The force through the crutch was recorded using a set of scales to ensured consistency between measurements.

Results: The main foot pressure characteristics are listed below. The table shows the mean value of the normal parameter, along with the percentage difference of the mean values when the crutch was used. The results of a paired-t test are also given to indicate the statistical significance of the change.

In addition to the above quantitative changes, qualitative assessment of the data showed an alteration in the loading pattern with reduced push off forces. However, eighteen of the twenty feet showed no alteration in the pattern of pressure distribution.

Conclusion: These results show that a crutch affects normal gait by generally increasing the time parameters and reducing the load and pressure parameters. The only parameters not significantly affected by the crutch were “Push off duration” and “Time to peak heel strike pressure”.

The aim of this study was to investigate if the rotational axis of normal human shoulders moves during flexion in the sagittal plane.

Twenty four shoulders were measured in twelve normal volunteers, aged 25-42, height range 1.65-1.88 m and weight range 63–120 Kg. Each subject had surface markers placed on their iliac crests, mastoid processes and upper arms. Joint movement was video recorded as shoulders were actively flexed and extended in the sagittal plane. For each joint, a typical flexion sweep was selected and replayed into a computerised imaging system, where still frames were captured at 20 degree intervals from 20 to 120 degrees. These images were analysed to extract the co-ordinates of each marker. The coordinates were then processed to determine the Instant Centres of Rotation (ICR) for each angle of flexion. These ICR’s were then plotted to derive the Rotational Axis Pathway (RAP) for each shoulder joint.

The results indicate that throughout the flexion arc, the rotational axis is located in the region of the humeral head. At the start of the arc the rotational axis is in the anterio-superior part of the shoulder joint. As the shoulder flexes forward the rotational axis moves posteriorly following a curved pathway. In 18 cases the RAPs moved posterio-inferiorly and in six cases the RAPs moved posterio-superiorly. The pathways can be quantified in terms of their curved pathway lengths and the displacements of their end points from their start points. In the case of the 18 posterio-inferior pathways, the mean pathway length was 98.3 mm (SD=31.5) and the mean posterior/inferior displacements were 59.6 mm (SD=34.7) and 43.2 mm (SD=24.6) respectively. In the case of the 6 posterior-superior pathways, the mean pathway length was 109.4 mm (SD=40.2) and the mean posterior/ superior displacements were 94.7 mm (SD=43.9) & 20.9 mm (SD=11.1) respectively. The variation in inferior-superior displacement of the axis may be due to normal variations in scapula movement during forward flexion.

This investigation indicates that in normal subjects, the rotational axis moves posteriorly during flexion.