Opening wedge high tibial osteotomy is an attractive surgical option for physically active patients with early osteoarthritis and varus malalignment. Unfortunately use of this surgical technique is frequently accompanied by an unintended increase in the posterior tibial slope, resulting in anterior tibial translation, and consequent altered knee kinematics and cartilage loading(1). To address this unintended consequence, it has been recommended that the relative opening of the anteromedial and posterolateral corners of the osteotomy are calculated pre-operatively using trigonometry (1). This calculation assumes that the saw-cut is made parallel to the native posterior slope; yet given the current reliance on 2D images and the ‘surgeon's eye’ to guide the saw-cut, this assumption is questionable. The aim of this study was to explore how accurately the native posterior tibial slope is reproduced with a traditional freehand osteotomy saw-cut, and whether novel 3D printed patient-specific guides improve this accuracy. 26 fourth year medical students with no prior experience of performing an osteotomy were asked to perform two osteotomy saw-cuts in foam cortical shell tibiae; one freehand, and one with a 3D printed surgical guide (Embody, London) that was designed using a CT scan of the bone model. The students were instructed to aim for parallelity with a hinge pin which had been inserted (with the use of a highly conforming 3D printed guide) parallel to the posterior slope of the native joint. For the purpose of analysis, the sawbones were consistently orientated along their mechanical and anatomical tibial axes using custom moulded supports. Digital photographs taken in the plane of the osteotomy were analysed with ImageJ software to calculate the angular difference in the sagittal plane between the hinge-pin and saw-cut. Statistical analysis was performed with SPSS v21 (Chicago, Illinois); a paired t-test was used to compare the freehand and patient-specific guide techniques. Statistical significance was set at a p-value <0.05.Introduction
Methods
Large heads offer substantial advantages over small ones in hip arthroplasty, as they are far less likely to dislocate. This feature is of particular benefit in very dysplastic females who often have a degree of joint laxity making dislocation a real possibility. Large metal heads have a range of problems, so registries report that they are now being substituted by large ceramic heads, typically reducing in diameter by 15% or more from the native size. All current designs of the femoral ball heads, whether for resurfacing of replacement share a unique design characteristic: a subtended angle of 120° defining the proportion of a sphere that the head represents. A novel design has recently been proposed that might reduce conflict between the femoral ball head rim and the iliopsoas tendon. This paper explains the problem of iliopsoas impingement on femoral heads of native diameter, and the consequences. Using MRI, we measured the contact area of the Iliopsoas tendon on the femoral head in sagittal reconstruction of 20 hips with symptoms of FAI. We also measured the Articular extent of the femoral head on 40 normal hips and 10 dysplastic hips. We then performed virtual hip resurfacing on normal and dysplastic type hips, attempting to avoid the overhang of the rim inferomedially. The contact area of the Iliopsoas tendon on the femoral head in extension is well visualized (Figure 1). The femoral head articular surface has a subtended angle of 120° anteriorly and posteriorly, but only of 100° medially. Virtual surgery in a femoral head of a dysplastic hip showed that when the femoral head is resurfaced with an anatomic sized component, the femoral ball head has a 20° skirt of metal protruding medially where iliopsoas articulates (figure 2). Reducing this by 15%, (eg to put a 40mm ball head onto a hip that had a 46mm femoral diameter), completely avoids any chance of iliopsoas tendon using the femoral head as a fulcrum. MRI of a dysplastic hip with a 40mm ball shows that iliopsoas impingement is hard to substantiate (figure 3).Material and Methods
Results
The Oxford Hip Score (OHS), the Harris Hip Score (HHS) and WOMAC are examples of patient reported outcome measures (PROMs) have well documented ceiling effects, with many patients clustered close to full marks following arthroplasty. Any arthroplasty that offers superior function would therefore fail to be detectable using these metrics. Two recent well conducted randomised clinical trials made exactly this error, by using OHS and WOMAC to detect a differences in outcome between hip resurfacing and hip arthroplasty despite published data already showing in single arm studies that these two procedures score close to full marks using both PROMS. We had observed that patients with hip resurfacing arthroplasty (HRA) were able to walk faster and with more normal stride length than patients with well performing hip replacements, but that these objective differences in gait were not captured by PROMs. In an attempt to capture these differences, we developed a patient centred outcome measure (PCOM) using a method developed by Philip Noble's group. This allows patients to select the functions that matter to them personally against which the success of their own operation will be measured. Our null hypothesis was that this PCOM would be no more successful than the OHS in discriminating between types of hip arthroplasty. 22 patients with a well performing Hip Resurfacing Arthroplasty were identified. These were closely matched by age, sex, BMI, height, preop diagnosis with 22 patients with a well performing conventional THA. Both were compared with healthy controls using the novel PCOM and in a gait lab. PROMs for the two groups were similar, while HRA scored higher in the PCOM. The 9% difference was significant (p<0.05). At top walking speed, HRA were 10% faster, with a 9% longer stride length. Outcome measures should be able to detect differences that are clinically relevant to patients and their surgeons. The currently used hip scores are not capable of delivering this distinction, and assume that most hip replacements are effectively perfect. While the function of hip replacements is indeed very good, with satisfaction rates high, objective measures of function are essential for innovators who are trying to deliver improved functional outcome. The 9% difference in PCOM found in this small study reflects the higher activity levels reported by many, and of similar magnitude to the 10% difference in top walking speed, despite no detectable difference in conventional PROMS. PCOMs may offer further insight into differences in function. For investigators who wish to develop improvements to hip arthroplasty, PCOMs and objective measures of gait may describe differences that matter more to patients than conventional hip scores.Results
Discussion
Hip impingement causes clinical problems for both the native hip, where labral or chondral damage can cause severe pain, and in the replaced hip, where subluxation can cause squeaking/metallosis through edge loading, or can cause dislocation. There is much research into bony/prosthetic hard impingements showing that anatomical variation/component mal-positioning can increase the risk of impingement. However, there is a lack of basic science describing the role of the hip capsule and its intertwined ligaments in restraining range of motion, ROM, and so it is unclear if careful preservation/repair of the capsular ligaments would offer clinical benefits to young adults, or could also help prevent edge loading in addition to reducing the postoperative dislocation rate in older adults. This in-vitro study quantifies the ROM where the capsule passively stabilises the hip and compares this to hip kinematics during daily activities at risk for hip subluxation. Ten cadaveric left hips were skeletonised preserving the joint capsule and mounted in a testing rig that allowed application of loads, torques and rotations in all six-degrees of freedom (Figure 1). At 27 positions encompassing a complete hip ROM, the passive rotation resistance of each hip was recorded. The gradient of the torque-rotation profiles was used to quantify where the capsule is taut/slack and after resecting the capsule, where labral impingement occur. The ROM measurements were compared against hip kinematics from daily activities. The capsule tightly restrains the hip in full flexion/extension with large slack regions in mid-flexion. Whilst ligament recruitment varies throughout hip ROM, the magnitude of restraint provided is constant (0.82 ± 0.31 Nm/degree). This restraint acts to prevent or reduce loading of the labrum in the native hip (Figure 2). The measured passive rotational stability envelope is less than clinical ROM measurements indicating the capsule does provide restraint to the joint within a relevant ROM. Activities such as pivoting, stooping, shoe tying and rolling over in bed all would recruit the capsular ligaments in a stabilising role. The fine-tuned anatomy of the hip capsule provides a consistent contribution to hip rotational restraint within a functionally relevant ROM for normal activities protecting the hip against impingement. Capsulotomy should be kept to a minimum and routinely repaired in the native hip to maintain natural hip mechanics. Restoring its native function following hip replacement surgery may provide a method to prevent subluxation and edge loading in the replaced hip.
Our primary hypothesis was simple: does gait on a downhill gradient distinguish between types of knee arthroplasty? Our secondary hypotheses were these: are stride length and other kinematic variables affected by cruciate ligament integrity following knee arthroplasty? Ethical approval was sought and gained prior to commencement of the study. 52 subjects were tested on the instrumented treadmill, 3 groups (UKA, TKA, and young healthy control) of 19, 14, and 19 respectively. The two high performing arthroplasty groups were recruited from a database of patient related outcome measures (PROMs) and were chosen based on high Oxford knee scores (OKS) with a minimum 12 months post hip arthroplasty.Introduction
Participants
Outcome measures are an essential element of our industry: comparing a novel procedure against an established one requires a reliable set of metrics that are comprehensible to both the technologist and the layman. We surmised that a detailed assessment of function before and after knee arthroplasty, combined with a detailed set of personal goals would enable us to test the hypothesis that less invasive joint and ligament preserving operations could be demonstrated to be more successful, and cost effective. We asked the simple question: how well can people walk following arthroplasty, and can we measure this? Using a treadmill, instrumented with force plates, we developed a regime of walking at increasing speeds and on varying inclines, both up and down hill. The data from the force plates was then extracted directly, without using the proprietary software that filtered it. Code was written in matlab script to ensure that missed steps were not mistakenly attributed to the wrong leg, automatically downloading of all the gait data at all speeds and inclines. The pattern of gait of both legs could then be compared over a range of activities. Wide variation is seen in gait both before and after arthroplasty. The variables that are easiest to explain are these:
width of gait – this appears to be a pre-morbid variable, not easily correctible with surgery. (figure 1) top walking speed – total knee replacement is associated with 11% lower top speeds than uni knees or normals (p < 0.05) change in stride length with increasing speed: normal people increase their walking speed by increasing both their cadence and their stride length incrementally until a top stride length is reached. Patients with a total knee replacement do not increase their stride length at a normal rate, having to rely on increasing cadence to deliver speed increase. Patients with uni or bi-compartmental knee replacements increase speed like normal people. Downhill gait: as many as 40% of fit patients with ‘well functioning’ total knee replacements choose not to walk downhill at all, while all fit patients with ‘well functioning’ partial replacements are able to do this. Those who can manage, can only manage 90% of the normal speed, unlike unis which are indistinguishable from normal (p < 0.05)Materials and methods
Results
The 3D shape of the normal proximal femur is poorly described in current designs of proximal femur prosthesis. Research has shown that in current implant designs with small diameter femoral heads the moment arm of the ilio-psoas tendon is reduced causing weakness in full extension, while large femoral heads cause psoas tendon impingement on the femoral head neck junction [1]. The femoral head-neck junction thus directly influences the hip flexor muscles' moment arm. Mathematical modeling of proximal femoral geometry allowed a novel proximal femur prosthesis to be developed that takes into account native anatomical parameters. We hypothesized that it is possible to fit a quadratic surface (e.g. sphere, cylinder…) or combinations of them on different bone surfaces with a relatively good fit. Forty six ‘normal’ hips with no known hip pathology were segmented from CT data. Previous research has shown the femoral head to have a spherical shape [2], the focus here was therefore mainly on the neck. The custom-written minimization algorithm, using least squares approximation methods, was used to optimize the position and characteristics of the quadratic surface so that the sum of distances between a set of points on the femoral neck and the quadratic surface was minimized. Furthermore, to improve upon current design regarding the transition between head and the neck, we recorded the position of the head neck articular margin in addition the slope of the transition from head to neck in the above 46 hips.INTRODUCTION:
METHODS:
A recent PRCT failed to demonstrate superiority of HRA over THA at low speeds. Having seen HRA walk much faster, we wondered if faster walking speed might reveal larger differences. We therefore asked two simple questions: Does fast or uphill walking have an effect on the observed difference in gait between limbs implanted with one HRA and one THA? If there is a difference in gait between HRA and THA implanted legs, which is more normal?
There were 3 females and 6 males in the study group, who had a mean age of 67 (55–76) vs the control group 64 (53–82, p = 0.52). The BMIs of the two groups did not differ significantly (28 v 25, p = 0.11). The mean average oxford score of included patients was 44 (36–48). Radiographs of all subjects were examined to ensure that implanted components were well fixed. The mean time from THA operation to gait assessment was 4 years (1–17 yrs) and that for HRA was 6 years (0.7–10 yrs, p = 0.31). Subjects in this study had a mean TWS of 6.8 km/hr (5–9.5), and a mean TWI of 19 degrees (10–25 degrees).INTRODUCTION
METHODS
Varus alignment of the knee is common in patients undergoing unicondylar knee replacement. To 56 patients (31 varus, 25 normal) were evaluated through CT imaging. Images were segmented to create 3D models and aligned to a tri-spherical plane (centres of spheres fitted to the femoral head and the medial and lateral flexion facets). 30 key co-ordinates were recorded per specimen to define the important axes, angles and shapes (e.g. spheres to define flexion and extension facet surfaces) that describe the femoral condylar geometry using in-house software. The points were then projected in sagittal, coronal and transverse planes. Standardised distance and angular measurements were then carried out between the points and the differences between the morphology of normal and varus knee summarised. For the varus knee group, trends were investigated that could be related to the magnitude of varus deformity.Introduction:
Methodology:
It has been proposed that higher knee adduction moments and associated malalignment in subjects with severe medial knee joint osteoarthritis (OA) is due to anatomical deformities as a result of OA [1, 2]. The emergence of patient-matched implants should allow for correction of any existing malalignment. Currently the plans for such surgeries are often based on three dimensional supine computed tomography (CT) scans or magnetic resonance imaging (MRI), which may not be representative of malalignment during functional loading. We investigated differences in frontal plane alignment in control subjects and subjects with severe knee joint OA who had undergone both supine imaging and gait analysis. Fifteen subjects with severe knee OA, affecting either the medial or lateral compartment, and 18 control subjects were selected from a database established as part of a larger study. All subjects had undergone gait analysis using the Vicon motion capture system. OA subjects had undergone routine CT scans and were scheduled for knee joint replacement surgery. Control subjects had no known musculoskeletal conditions and had undergone MRI imaging of hip, knee and ankle joints. Frontal plane knee joint angles were measured from supine imaging (supine) and from motion capture during standing (static) and during gait at the first peak ground reaction force (gait). OA subjects had a significantly higher BMI (p < 0.01) and different gender composition (13 males and 2 females vs 4 males and 5 females; p = 0.03) compared with controls. Multiple linear regression analysis indicated no significant confounding effect of these differences on frontal plane angles measured in supine, static or gait conditions. For both OA and healthy subjects, frontal plane knee angles were significantly higher during gait compared with supine (p = 0.03 and 0.02, respectively). There were also significant differences in knee alignment between OA and healthy subjects for supine and static (p < 0.05) but not for gait, although this was approaching significance (p = 0.052). Overall there seemed to be higher variation in alignment in the OA subjects (Fig. 1). The significantly higher frontal plane knee joint angles measured in both control and OA subjects during gait compared with supine imaging indicate that functional alignment should be taken into consideration when planning patient-specific surgeries. Higher variation in OA patients may be due to alterations in gait patterns due to pain or degree of wear in their osteoarthritic joints, and requires further investigation. In addition, methodological considerations should be taken when comparing alignment from measurements taken with imaging and motion capture to avoid systematic errors in the data. In conclusion, we believe that both supine and loadbearing imaging are insufficient to gain a full representation of functional alignment, and analysis of functional alignment should be routinely performed for optimal surgical planning.
Cam type femoroacetabular impingement (FAI) may lead to osteoarthritis (OA)[1]. In 2D studies, an alpha angle greater than 55° was considered abnormal however limitations of 2D alpha angle measurement have led to the development of 3D methods [2–4]. Failure to completely address the bony impingement lesions during surgery has been the most common reason for unsuccessful hip arthroscopy surgery [5]. Robotic technology has facilitated more accurate surgery in comparison to the conventional means. In this study we aim to assess the potential application of robotic technology in dealing with this technically challenging procedure of cam sculpting surgery. CT scans of three patients' hips with severe cam deformity (A, B and C models) were obtained and used to construct 3D dry bone models. A 3D surgical plan was made in custom written software. Each 3D plan was imported into the Acrobot Sculptor robot and bone resection was carried out. In total, 42 femoral models were sculpted (14/subset), thirty of which were performed by a single operator and the remaining 12 femurs were resected by two other operators. CT of the pre/post resected specimens was segmented and a 3D alpha angle and head neck ratios were measured [3–4] and compared using Mann-Whitney U test. Coefficient of variation (CV) was used to determine the degree of variation between the mean and maximum observed alpha angles for inter and intra observer repeatability.Introduction:
Methods:
Barriers to the adoption of unicompartmental knee arthroplasty (UKA) by new consultants could be explained by its higher revision rate, to which mal-positioned components contribute. The aim of this study was to determine whether robotic technology enables inexperienced surgeons to perform accurate UKAs when compared to current conventional methods After randomisation, sixteen trainees who had never performed UKAs performed three medial UKAs (Corin Uniglide), one per week, on dry-bone simulators by either robotic (Sculptor RGA) or conventional methods. They were instructed to match a universal 3D-CT based pre-operative plan that would result from a UKA based on the conventional jigs and operating guide. The knees were laser scanned and software used to compare the planned and actual implant positions. Feedback was given to trainees between attempts. Translational and rotational positioning errors were measured in all six degrees of freedom for both components At all attempts robotic medial UKAs were more accurate in both translational and rotational alignments for both components reaching statistical significance (p<0.005) at all attempts for rotational errors. Considering outliers, the maximum rotational errors of the robot group was 9° and 7° for the tibial and femoral components respectively. For the conventional group this reached 18° and 16° for the tibial and femoral components respectively Robotic technology allows inexperienced surgeons to perform medial UKAs on dry bone models with acceptable accuracy and precision on their first attempt. Conventional jigs do not. The adoption of robotic technology might provide new consultants with the confidence to offer UKAs to their patients by limiting the inaccuracies inherent in conventional equipment.
Osteoarticular loss in a major weightbearing joint is one of the many consequences of military conflict. While minor in terms of life and limb salvage, when rehabilitation is being planned, a small amount of joint damage can make a large impact on the level of long term disability. Reconstruction methods include allograft, massive replacement, arthrodesis and amputation. We have been developing a suite of technologies that contribute to the reconstruction of such injuries including assessment of disability in a fully instrumented gait lab, modelling of the injury using low dose CT, analysis of the extent of loss and creation of stereolithograph files of the bones, planning of the surgical procedure including implants as needed, custom manufacture of osteotomy guides and prostheses if required and technology assisted surgery, including active constraint robots. We report 3 cases of soldiers who have suffered osteoarticular loss to part of the knee, two from high velocity rounds and one from an IED. All 3 have received custom partial knee replacements preserving their cruciates, the other compartment and the patella-femoral joint. No major technical issues have been encountered. The surgery is quick and recovery simple, with the prospect of normal painfree pedestrian life. Exchange of the bearing will be necessary. It is expected and planned for once a decade as a day case procedure. While the surgery is expensive, it is highly cost-effective, as it restores near normal gait. It does not appear to be particularly risky, as the volume of tissue being excised is small, and highly constrained. The same approach is now being developed for deployment in civilian life for post traumatic osteoarthritis.
The introduction of the Stanmore Implants Savile Row mobile-bearing UKA procedure in July 2011 marked a world first – the use of a patient-specific knee implanted with robotic technology – the Sculptor Robotic Guidance Arm (RGA). This union gives a truly personalised solution by designing an implant for each patient based upon preoperative CT data and using Sculptor RGA to prepare the bone accurately so that the implant is correctly positioned as planned. The purpose of this study is to evaluate the accuracy of Sculptor RGA both in-vitro and in-vivo. We report on the accuracy of our first clinical procedures. In-vitro: CTs of plastic-bones were used to create plans for Sculptor RGA, establishing a relationship between the implant position and plastic-bone (planned-transform). Sculptor RGA was then used to prepare bones for 16 UKA implants mimicking the clinical set-up. The implants were placed in the prepared bones without cement. A coordinate-measuring-arm was used to register a)the bone, and b)the implant in relation to the bone (achieved-transform). The difference between planned-and-achieved transforms gives the error in implant position. In-vivo: Preoperative CTs of 8 OA patients, acquired using the low-dose Imperial Knee CT protocol, were used to plan the position and the shape of the patient-specific implants. Intra-operatively, Sculptor RGA was used to register and prepare the bone and the implants were cemented in place. Post-operative CTs were also acquired. Two techniques were used to measure planned-to-achieved positions of the implants: 1). Preoperative-to-postoperative CT image registration followed by extraction of the achieved implant position and comparison with the plan, 2). Surface-to-surface registration of bone-models segmented from the preoperative and postoperative CTs followed by extraction of the achieved implant position and comparison with the plan.Introduction
Methods
High functional aspirations and an active ageing population equate to a growing number of patients awaiting hip arthroplasty demanding superior biomechanical function. The purpose of this study was to compare the biomechanics of top walking speed between two commonly used hip arthroplasty procedures to determine if a performance advantage existed. A retrospective comparative study was performed using sixty-seven subjects, twenty-two subjects in both hip resurfacing and total hip arthroplasty groups along with twenty-three healthy controls. All arthroplasty subjects were recruited based on high psychometric scoring and had been performed through a posterior approach, and had been discharged from follow-up. On an instrumented treadmill each subject was measured by a researcher blinded to which procedure that patient had undergone. After a six minute acclimatization period, the speed was increased incrementally until top walking performance had been attained. At all increments, ground reaction forces and temporospatial measurements were collected.Background
Methods
The combination of patient-specific “just-in-time” implant manufacture and robotic technology has not yet been reported. The robot enables accurate placement of anatomically-matched implants. It should be cost-effective, simplify the procedure, and reduce instrumentation. The aims of this study were to determine whether the procedure was safe, radiographically accurate, and comparable in time and cost to conventional arthroplasty. All patients over 3 months post-op were included. Component position, orientation and size were determined from CT scans by the surgeon prior to manufacture. The implants were inserted using the Sculptor robot, which is supplied free of cost (Savile Row, Stanmore Implants, UK). Following registration, bone was milled away using a high-speed burr under haptic control of the robot. The implants were cemented and a mobile bearing inserted. Patients were followed up clinically and radiographically. Oxford and EQ-5D scores were obtained. Costs of the implant, instruments, and consumables were calculated and compared to published data for conventional UKA and TKA.Background
Methods
Are there really ‘conventional’ bearings, offering more security and less risk than the ‘alternative’ bearings that feature in the programme? Alternative, when used as an adjective has 2 meanings: offering or expressing a choice, as in several alternative plans different from or functioning outside the usual or conventional: eg alternative newspaper, alternative rock music, alternative medicine This paper reviews the elements that make up the bearing couples available today in the developed world, and tests each bearing against these meanings. what are the alternatives? The materials available today fall into the following broad families:Introduction
Materials
Metal on Metal hip resurfacing (MoM HR) can be an effective operation for the young arthritic hip population. However, errors in cup orientation have been associated with increased wear, circulating blood metal ions, and soft tissue abnormalities that can lead to premature failure of the bearing surface and subsequent revision surgery. While image free computer guidance has been shown to increase surgical accuracy in total hip arthroplasty, the role of image based technology in MoM HR is unclear. In this study, we compared the accuracy of cup orientation in MoM HR performed by either freehand technique or CT based navigation. Seventy five patients (81 hips) underwent either freehand (n=42) or navigation (n=39) surgery, both requiring a three dimensional (3D) CT surgical plan. Surgery was conducted by hip specialists blind to the method of cup implantation until the operation. Deviation in inclination and version from the planned orientation, as well as, number of cups within a 10° safe zone and 5° optimal zone of the target position was calculated using post operative 3D CT analysis. Error in inclination was significantly reduced with navigation compared to freehand technique (4° vs 6°, p=0.02). We could not detect a difference between the two groups for version error (5° vs 7°, p=0.06). There was a significantly greater number of hips within a 10° (87% vs 67%, p=0.04) and 5° (50% vs 20%, p=0.06) safe zone when navigated. Image based navigation can substantially improve accuracy in cup orientation. The results of our freehand group appear better than historic controls, suggesting the use of a 3D plan may help to reduce technical error and improve the learning curve in this technically demanding procedure. We advocate the use of image based navigation in MoM hip resurfacing arthroplasty.
The current generation of knee replacements are based upon assumptions from kinematic studies that preceded their designs. These implants were further limited by practical restrictions imposed by affordability, materials and manufacturing, and finally by the methods available to surgeons to prepare the bone and implant them. The early designs of knee seldom distinguished left from right, as the early kinematic work had not appreciated the very different functions of the medial and lateral compartments. Trochlea shape and position within devices was also limited by the published work on the way the knee bends. Surgical insertion has been limited to landmark based registration, and adjustment of the kinematics by soft tissue releases. However accurately such operations were performed, they could not restore normal function, as the kinematics of the joint were quite different from the normal knee. Recently, we have begun to appreciate three distinct axes of the knee joint: the flexion axis, the extension axis and the trochlea axis. These can be reliably found from 3d imaging, but cannot be immediately established by eye, or by conventional jigs, which must rely on unreliable landmarks acquired in surgery. The current market leaders in knee joint sales do not reflect these three axes in their joint designs, so the instrumentation used to insert them cannot restore the kinematics of the normal knee. The emerging partial replacements can be designed to take the axes and their resulting kinematics into account. If they are then inserted using robotic assistance, or patient specific guides, they can restore joints to these axes reliably. Knee function following such conservative surgery reflects this improvement in kinematics with higher functional scores and faster top walking speeds than has ever been possible using conventional devices inserted using the conventional landmark based surgical techniques.
The radiographic analysis of over 5000 metal on metal (MoM) hips using Ein Bild Roentgen Analyse (EBRA) software have been recently published in an attempt to determine the influence of cup orientation on bearing function. The validation of this software relies one study, conducted in a phantom pelvis without a femoral head in situ. Three dimensional computed tomographic (3D-CT) has been shown to be more accurate for hip and knee arthroplasty than plain radiographs for measurements of component orientation and position. The accuracy of EBRA when compared to 3D-CT for MoM hips specifically is unclear. We measured the cup orientation of 96 large diameter MoM hips using EBRA analysis of plain radiographs and compared this to 3D-CT. All measurements were made using the radiographic definition of cup orientation. The mean difference in version between the two imaging modalities was 8°; with wide limits of agreement of −21.2° and +5.6°. Three retroverted cups were not detected by EBRA. The mean difference in inclination values was 0.3°, but there was up to 9°difference between imaging modalities. When measured by 3D CT, 64% of hips were within a 10° safe zone around 45° inclination and 20° version, compared to only 24% when measured by EBRA (Fishers Exact test, p< 0.0001). The measurement of cup orientation of MoM hips using EBRA software is insufficiently accurate, particularly for the assessment of cup version. The cup rim is obscured by the large diameter femoral head on plain radiographs. Research studies using EBRA analysis for version have limited value if accuracy of more than 20 degrees is required to draw conclusions. This software may not be suitable to measure the performance of a device or surgeon. The limitations of EBRA can be overcome, if 3D-CT with an extended Hounsfield scale for data capture is used.
