Abstract

Open cheilectomy is an established surgical treatment for hallux rigidus. Cheilectomy is now being performed using minimally invasive (MIS) techniques. In this prospective study we report the outcome of minimally invasive cheilectomy comparing the results with a matched group who had cheilectomy using standard open procedure

Introduction

Symptomatic hallux valgus is a common clinical problem, current trends is towards minimally invasive procedures. The goal of this study is to evaluate the outcome of minimally invasive chevron ostoetomy, comparing it with a matched group who had open chevron osteotomy.

Open cheilectomy is an established surgical treatment for hallux rigidus. Cheilectomy is now being performed using minimally invasive(MIS) techniques. In this prospective study we report the outcome of minimally invasive cheilectomy comparing the results with a matched group who had cheilectomy using standard open procedure.

Prospective study of 47 patients. 22 patients had MIS cheilectomy between March 2009 and September 2010. We compared the outcome with a matched group (25 patients) who had open cheilectomy. Functional outcome was assessed using the Manchester Oxford Foot and ankle questionnaire (MOXFQ). The MOXFQ is a validated 16-item, patient-generated questionnaire designed to be self-completed and used as an outcome measure for foot surgery. It comprises three domains foot pain, walking and standing problems and social interaction. Total score ranges from 0 (best score) to 64 (worst score). Patients completed preoperative and postoperative questionnaires. Patients’ satisfaction and complications were recorded.

In the MIS group, the median follow up was 11 months (4–23. The median preoperative MOXFQ score was 34/64(23) and the median postoperative score was 19/64 (p value <0.02) In the open group the median follow up was 17 months (9–27). The median preoperative MOXFQ score was 35/64 and the median postoperative score was 7.5/64 and this difference was statistically significant (<0.0001). There were three failures in the open group (Fusion) compared to none in the MIS.

MIS cheilectomy is an effective alternative procedure with satisfactory functional outcome and high patient satisfaction. Results are comparable to the standard open cheilectomy with a lower apparent failure rate. The results of our randomised controlled trial comparing MIS cheilectomy to open cheilectomy are awaited.

Background

Morton's Neuroma is a proximal neuralgia that affects the web spaces of the toes and is currently of unknown aetiology. Currently surgical excision is considered the gold standard treatment based on RCT and cohort studies. However patient derived outcomes have not previously been assessed. We addressed these aspects with our study.

The anterior femoral cortical line (AFCL) is an anatomical landmark which has been used by the senior author for 20 years to assess femoral rotation in over 4000 TKR's. The AFCL describes the alignment of the anterior cortex of the distal femur proximal to the trochlear articular cartilage.

The AFCL was compared with the surgical epicondylar (SEA), anteroposterior (Whiteside's line) and posterior condylar (PC) axes using 50 dry-bone cadaveric femora, 16 wet cadaveric specimens, 50 axial MRI's and 58 TKR patients intra-operatively.

In the dry-bone/cadaveric femora (measuring relative to the SEA the AFCL and Whiteside's AP axis were 1° externally rotated and the PC axis was 1° internally rotated. By MRI (relative to the SEA) the AFCL was 8° internally rotated, Whiteside's was 2° externally rotated and the PC axis was 3° internally rotated. In the clinical study (measuring relative to a perpendicular to Whiteside's line alone) the AFCL was 4° degrees internally rotated, which equates to 2-3° of internal rotation relative to the SEA.

The AFCL is another axis, completing the ‘compass points’ around the knee. It may prove particularly useful when one or all of the other reference axes are disturbed such as in revision TKR, lateral condylar hypoplasia or where there has been previous epicondylar trauma. We suggest building in 5° external rotation with respect to the anterior femoral cortical line for femoral component rotation.

Introduction: Although the primary aim of Total Knee Replacement (TKR) is to relieve chronic joint pain, 10–20% of patients experience unexplained chronic pain after surgery. One possible cause of this pain is central sensitisation. Prolonged exposure to a noxious input can lead the central nervous system to become sensitised to pain (central sensitisation), which can become self-sustaining and persist after the removal of the noxious stimuli i.e after TKR. The aim of this study was to determine if knee osteoarthritis (OA) patients awaiting TKR have evidence of sensory perception abnormalities, by comparing detection and pain thresholds from OA patients to those of age- and gender-matched healthy participants.

Patients and Methods: Quantitative Sensory Testing (QST) was performed on 107 knee OA patients on the waiting list for primary TKR and 50 age- and gender-matched healthy participants without knee pain or TKR. QST assesses somatosensory function through measuring participant responses to external stimuli of controlled intensity. QST was performed on both knees and the pain-free forearm of all participants. Von Frey filaments were used to measure touch detection thresholds, a digital Algometer to measure pressure pain thresholds, and the MSA Thermotest to measure detection and pain thresholds to hot and cold. Significant differences in the median threshold values between knee OA patients and healthy participants were tested for using Mann-Whitney U tests.

Results: Detection thresholds: OA patients had significantly higher detection thresholds for hot and cold (both p< 0.05) in the index knee (but not at other sites) compared to healthy participants. Touch detection thresholds were significantly higher at all body sites in OA patients compared to healthy participants (all p< 0.001).

Pain thresholds: Pressure pain thresholds were significantly lower in OA patients at all body sites (all p< 0.001) but there were no significant differences in hot or cold pain thresholds between OA patients and healthy participants at any body site (all p> 0.05).

Discussion: This study showed that knee OA patients have modality-specific sensory and pain perception abnormalities. These included thermal hypoesthesia (reduced sensitivity) in the index knee and tactile hypoesthesia at all body sites tested, alongside hyperalgesia (increased pain sensitivity) to pressure pain at all body sites. Future research aims to determine if these sensory perception abnormalities are predictive of chronic pain after TKR.

The main surgical options for hallux rigidus include dorsal cheilectomy, arthroplasty and fusion. Compared to other surgical procedures dorsal cheilectomy involves relatively less bone loss, maintains joint motion and still leaves the potential for a further salvage procedure such as arthrodesis. The Manchester – Oxford Foot and Ankle Questionnaire (MOXFQ) has been endorsed by the British Foot and Ankle Society and recommended as a tool to measure surgical outcome.

The aim of this study was to prospectively follow-up patients following dorsal cheilectomy for hallux rigidus using the patient-reported outcome measure, the MOXFQ. Patients were deemed suitable for dorsal cheilectomy if they had symptoms of dorsal impingement (painful restriction of terminal dorsiflexion of the hallux), with absence of pain in the mid-range of passive movement (indicative of end stage arthritis), and dorsal osteophytes on radiographs.

Twenty five patients with a mean age of 62 years (range: 39 – 80 years) and comprising 17 females underwent dorsal cheilectomy for hallux rigidus. At a mean follow-up of 17 months (range 9–27 months), the mean MOXFQ score improved from 33/64 (95% CI: 27 – 38) pre-operatively to 9/64 (95% CI: 6 – 13) post-operatively. Eighty-four percent of patients had a clinically significant improvement in their walking domain, 68% in the social domain, and 59% in the pain domain of the MOXFQ. Four patients were deemed failures: three patients underwent subsequent arthrodesis for persisting pain, and one patient had no improvement in any domain of the MOXFQ.

In conclusion, this study has provided further evidence of the success of dorsal cheilectomy as a reliable treatment option for hallux rigidus. It has also demonstrated the sensitivity of the patient focused reporting instrument the MOXFQ in assessing surgical outcomes in foot surgery.

Bone Cement Implantation Syndrome is a rare but serious complication following operations involving the use of cement for prosthesis fixation. The POSSUM scoring system has been shown to be a reliable predictor for morbidity and mortality in Orthopaedic surgery and a useful audit tool to observe effects of different treatment strategies. We have applied this scoring system to 6 consecutive patients that underwent cemented hemiarthroplasty for fractured neck of femur and subsequently died directly as a consequence of complications arising from cementation.

The average predicted mortality using the POSSUM score on the 6 patients that died was 28% (range 12–52%). This is markedly higher than the average POSSUM predicted mortality of fractured neck of femur patients in other studies (9.1–15.6%).

Our results indicate that patients who died as a direct result of cement implantation have a higher predicted mortality rate using the POSSUM score. The scoring system can be used to help identify fractured neck of femur patients that are susceptible to reaction to the cement. This allows the surgeon to consider alternative surgical options, such as cementless prostheses, during pre-operative planning.

The optimum design for the femoral component for cementless Total Hip Replacement is not known. We conducted an ethically approved, randomized and prospective trial to compare two radically different designs of fully hydroxyapatite (HA) coated femoral stems. We compared the original JRI Furlong stem with the Wright Anca fit stem which is more anatomical in design. The paper discusses the merits and disadvantages of these two stems. The same acetabular component was used in both samples. The only variable was the stem shape.

All patients placed on the senior authors’ waiting list for primary THR were asked if they would enter the trial. There were no restrictions for selection to the sample. Patients were then randomised for one of the two stems. All surgery was performed by or under the direct supervision of the senior author in one center. The surgical approach to the hip, Hardinge antero lateral, was the same in all cases. In all cases the same well-tried JRI CSF acetabular cup and bearings were used. Patients were x-rayed post operatively and reviewed and x-rayed at six weeks and then yearly.

360 patients had been entered into the trial, 219 females and 141 males. 203 patients had the JRI furlong hip implanted (56%) and 157 were in the Anca sample (44%). The periprosthetic and perioperative fracture rates for the two stems were found to be significantly different at three years into the study and the trial was stopped. The number of perioperative fractures in the Furlong group was 17(8.4%)and in the Anca sample 24 (15.3%). This is statistically significant. The possible reasons for this difference are discussed.

Anatomical fit and wedge shaped cementless stems require a more careful technique to avoid fracture during implantation and the shorter stemmed Anca hip proved less stable in the presence of a fracture necessitating further surgical intervention. It is still not certain whether the anatomical shape has long term advantages that may outweigh this initial disadvantage and the cohort of patients continue to be followed up.

Purpose of study: To determine whether positioning the operated limb in flexion on a CPM machine reduces blood loss following uncemented total knee replacement. We also sought to determine whether limb arterial pressure, represented by ankle-brachial pressure index (ABPI), is affected by position of the limb.

Methods: We compared two groups of age and sex matched patients undergoing uncemented total knee replacement. The control group were treated in the postoperative period with the leg positioned flat in extension (48 patients). The CPM group were placed on a CPM machine at 90° of knee flexion for twelve hours immediately post-operatively (46 patients). Haemoglobin and haematocrit levels, blood transfusion requirement, length of hospital stay, range of motion and complications were compared.

In addition, we measured ankle and brachial pressures in a separate group of 39 patients with the limb in three different positions: flat with the knee in extension (ABPI 1), raised with the knee in extension (ABPI 2), and finally with the knee flexed to 90° (ABPI 3).

Results: The CPM group demonstrated a smaller drop in haemoglobin level (2.5 vs 3.1, p=0.013) and a reduced blood transfusion requirement (3 patients vs 14 patients, p < 0.05) in the post-operative period when compared to the control group. There was no significant difference in haematocrit drop, discharge range of motion, complications or length of hospital stay between the two groups.

ABPI measurements were calculated in the standard fashion, the mean ABPI in each limb position being 1.17 (ABPI 1), 0.87 (ABPI 2) and 0.83 (ABPI 3) respectively.

Conclusions: Short-term use of a CPM machine in the post-operative period reduces blood loss and appears safe following uncemented total knee replacement. We found limb arterial pressure to be reduced to the greatest extent with the limb in a position of 90° flexion at the knee, which may account for the reduction in blood loss seen in the CPM group.

Background: The purpose of this study was to assess the anterior femoral cortical line (AFCL) as an additional anatomical landmark for determining intraoperative femoral component rotation in total knee arthroplasty. The AFCL was compared with the Epicondylar axis, the anteroposterior (AP) axis (Whiteside’s line), and the posterior condylar axis. Dry bone, cadaver, MRI and intra-operative measurements were compared.

Methods: Fifty dry bone femora, and 16 wet cadaveric specimens were assessed to identify the AFCL and this was compared against the 3 reference axes discussed above. Photographs were taken of the specimens with K-wires/marker pins secured to the reference axes and then a digital on-screen goniometer was used to determine the mean angular variations with respect to the Epipcondylar axis.

In the clinical trial, 58 consecutive patients undergoing total knee arthroplasty were included. After a routine exposure the AP axis was marked on each distal femur. The AFCL was then identified and the anterior femoral cortical cut was made parallel to this line. The angle between this cortical cut and the perpendicular to the AP axis was measured using a sterile goniometer.

In the MRI study, 50 axial knee images were assessed and the most appropriate slice/s determined in order to identify the AFCL and the other 3 reference axes and then their relationship was measured by an on-screen goniometer.

Results: In the cadaveric study the AFCL was a mean 1° externally rotated to the epicondylar axis (SD = 5°), White-side’s line was 1° externally rotated (SD = 4°) and the posterior condylar axis was 1° internally rotated (SD = 2°)

By MRI and with respect to the epicondylar axis, the AFCL was a mean 5° externally rotated (SD= 3), White-side’s Line was 1° externally rotated (SD = 2) and the posterior condylar axis was 3° internally rotated (SD = 2).

In the clinical study in 8 patients it was impossible to draw the AP axis because of dysplasia or destruction of the trochlea by osteoarthrosis. In the remainder the mean difference between the anterior femoral cortical line and Whiteside’s AP axis was 4.1 degrees internally rotated (SD = 3.8°). The lateral release rate for this cohort was 4%.

Conclusion: The anterior femoral cortical line provides an additional reference point, completing the ‘compass points’ around the knee. It has been shown in this study to be reliable in the laboratory, on MRI and in a clinical setting for assessing rotation of the femoral component. It may prove particularly useful when one or all of the other reference axes are disturbed such as in revision TKR, lateral condylar hypoplasia or where there has been previous epicondylar trauma.

Purpose: To assess the anterior femoral cortical line (AFCL) as an additional anatomical landmark for determining intraoperative femoral component rotation in total knee arthroplasty. The anterior femoral cortical line (AFCL) is an anatomical landmark which has been used by the senior author for 20 years to assess femoral rotation in over 4000 TKRs. The AFCL describes the alignment of the anterior cortex of the distal femur proximal to the trochlear articular cartilage.

Methods: The AFCL was compared with the surgical epicondylar axis (SEA), anteroposterior axis (Whiteside’s line) and posterior condylar (PC) axis using 50 dry-bone cadaveric femora, 16 wet cadaveric specimens, 50 axial MRI scans and 58 TKR patients intra-operatively.

Results: In the dry-bone and cadaveric femora (measuring relative to the SEA) the AFCL and Whiteside’s AP axis were 1° externally rotated and the PC axis was 1° internally rotated. With MRI (relative to the SEA) the AFCL was 8° internally rotated, Whiteside’s was 2° externally rotated and the PC axis was 3° internally rotated. In the clinical study (measuring relative to a perpendicular to Whiteside’s line alone) the AFCL was 4° degrees internally rotated, which equates to 2–3° of internal rotation relative to the SEA.

Conclusion: The AFCL is another axis, completing the ‘compass points’ around the knee. It may prove particularly useful when one or all of the other reference axes are disturbed such as in revision TKR, lateral condylar hypoplasia or where there has been previous epicondylar trauma. We suggest building in 5° external rotation with respect to the anterior femoral cortical line when judging femoral component rotation.

Aims: To determine the difference between macroscopic and microscopic appearances of the anterior cruciate ligament (ACL) in patients with osteoarthritis undergoing total knee replacement.

Methods: Patients admitted for routine total knee replacement (TKR) for osteoarthritis were assessed. The integrity of the ACL was noted as normal, moderately damaged (fissured) or complete rupture on a macroscopic level at the time of surgery. The ACL was sacrificed as a normal step in the operation and sent for histological analysis. The macroscopic and microscopic findings of ACL histology were compared using a common grading system (Grade 1 = normal, Grade 2 = moderately diseased, Grade 3 = severely diseased).

Results: The sample contained 48 patients, 17 male and 31 female, age range 55–87 years (mean 73). After exclusions, at surgery 7 ACL ruptures and 8 moderately diseased ligaments (defined by the presence of visible fissuring) were found despite negative pivot shift tests preoperatively. Of the 30 ACL’s that were found to be macroscopically normal, 22 of these (73%) were microscopically moderately or severely diseased.

Conclusion: We have found that a macroscopically normal ACL does not necessarily equate to microscopic integrity in the presence of osteoarthritis. This is an important consideration given the current trend towards unicompartmental knee replacement and highlights the possible need for investigation with appropriate imaging (MRI) and arthroscopy prior to this specific surgery.

Purpose: The aim was to determine whether intra-articular bupivacaine provides effective analgesia following ankle arthroscopy and whether it reduces the need for supplemental analgesia.

Methods: A power calculation revealed that 40 patients were required to provide a 5% significant level using a visual analogue scale. 40 patients were therefore randomised to receiving 20 mls of either bupivacaine or saline (control) after routine anterior ankle arthroscopic surgery. The tourniquet was released 10 minutes later. In recovery, supplementary analgesia of 2 tablets of co-codomol 30/500 orally or 50–100mg tramadol IV was available on request. A 10 day supply of 50mg diclofenac (8 hourly) and co-codomol 30/500 2 tablets (6 hourly) was provided.

A visual analogue scale (VAS) was employed as a direct indicator of pain and indirectly, supplemental analgesic requirement. Measurements were made pre-operatively and postoperatively.

Age, weight and tourniquet times were compared with Mann Whitney U test and Chi-square. Pain scores and analgesic requirements were compared using ANOVA at a 5% significance level.

Results: Pain scores were lower in the bupivacaine group compared to the control as was the need for supplemental analgesia.

Significance: We conclude that postoperative intraarticular bupivacaine provides effective analgesia following ankle arthroscopy.

The purpose of this study was to investigate the functional outcome of a group of patients following completely neglected tendo-achilles ruptures.

Between July 2001 and July 2002 we identified 6 patients who presented to the Foot and Ankle Service in Oxford with 7 chronic untreated complete ruptures of the tendo-achilles. There were 4 males and 2 females and the average age was 65 years (range 52 to 79). The average time since injury was 12.2 months (range 7 to 24). None of them had undergone any modality of treatment for this condition. From the history, a definite acute injury was confirmed in each patient. All patients had a palpable defect in the tendo-achilles between 4 and 8 cm from the insertion and the defects measured from 10 to 32 mm. In all case the Thompson test confirmed ongoing discontinuity and single leg heel raise was not possible on the affected side. Each patient was assessed using the scoring system of Leppilahti and concentric and eccentric power were assessed using the Kin-Com Dynamometer. The results indicate an average Leppilahti score of 65/100 with 1 excellent, 0 good, 3 fair and 2 poor. The isokinetic strength measurements demonstrated that plantar flexor power was on average 36% weaker than the normal side. These differences were most marked at the higher test speeds, which were on average 16% weaker than at the lowest test speed in the affected leg. Five out of 6 patients were pain free, with only one reporting mild pain. Objective testing demonstrated no differences in the range of movement between the injured and the normal side. All patients were satisfied with the outcome; however, most had some reservations, which related to ongoing weakness that prevented recreational activity.

Conclusion: At one year after injury the completely neglected tendo-achilles rupture in the older population is likely to be pain-free, to have full active ankle movement and to recover two-thirds of the power of plantar flexion compared to the unaffected limb. Ongoing weakness will prevent recreational activity but has minimal affect on activities of daily living including stair-climbing ability. We have established a benchmark of the natural history of this condition in the older patient against which the results of surgical treatment can be compared.

The intra-articular injection of local anaesthetic is frequently used for pain relief after arthroscopy. There is, however, no published evidence of the analgesic effect of bupivacaine in the ankle. In a randomised, double-blind study, 35 patients undergoing arthroscopy of the ankle were allocated to receive intra-articular saline or bupivacaine. Pain was assessed using pain scores and additional analgesic requirements. Intra-articular bupivacaine had a significant analgesic effect in the immediate post-operative period, reducing pain scores and the need for additional analgesics.

We recommend the use of intra-articular bupivacaine for post-operative analgesia in ankle surgery.

Introduction and Aims: The effects of injury to the posterior cruciate ligament (PCL) and posterior-lateral corner (PLC) on physical function are not as well documented compared to the more common injury to the anterior cruciate ligament. This study aimed at improving our understanding of PCL/PLC injury through gait analysis and electromyographic (EMG) testing.

Method: We studied 19 patients, average age 30 years (20–55) with clinically and radiologically confirmed PCL/PLC deficiency in isolation. Ninety percent of patients complained of instability when performing the activities of daily living and all complained of pain. All patients were assessed using the Lysholm and Gillquist functional knee score as well as gait analysis, including Kinematics, Kinetics and EMG of the quadriceps, hamstrings and gastrocnemius muscles. Findings were compared to our normal database. The mean Lysholm score was 51/100 (24–90). Those with a Lysholm greater than 50 were designated as ‘copers’.

Results: There were 12 ‘non-copers’ and seven ‘copers’. Fifty percent of patients demonstrated a varus thrust through stance. Forty-two percent of patients demonstrated hyperextension of the knee through stance. Sixty-three percent of patients demonstrated premature and prolonged hamstring activity. Thirty-seven percent of patients had premature activity of the gastrocnemius muscle in stance. Fifty-seven percent of the ‘copers’ demonstrated premature and prolonged hamstring activity through the gait cycle compared to forty-five percent of ‘non-copers’ (non-significant p=0.25 Fishers Exact Test). Fifty-five of ‘non-copers’ demonstrated premature activity of the gastrocnemius muscle in stance compared to none of the ‘copers’ (significant p=0.025 Fishers Exact Test).

Conclusion: The observed varus thrust may be responsible for the development of medial and patellofemoral compartment osteoarthritis, a recognised problem in PCL deficient knees. Hyperextension that occurs dynamically during gait could explain failure of PCL/PLC reconstruction over time. The observed abnormal hamstrings activity is unlikely to be a compensatory mechanism.

We analysed the histological findings in 1146 osteoarthritic femoral heads which would have been considered suitable for bone-bank donation to determine whether pathological lesions, other than osteoarthritis, were present. We found that 91 femoral heads (8%) showed evidence of disease. The most common conditions noted were chondrocalcinosis (63 cases), avascular necrosis (13), osteomas (6) and malignant tumours (one case of low-grade chondrosarcoma and two of well-differentiated lymphocytic lymphoma). There were two with metabolic bone disease (Paget’s disease and hyperparathyroid bone disease) and four with inflammatory (rheumatoid-like) arthritis. Our findings indicate that occult pathological conditions are common and it is recommended that histological examination of this regularly used source of bone allograft should be included as part of the screening protocol for bone-bank collection.

Introduction: Dysfunction of the patellofemoral mechanism presents as patella dislocation or subluxation with or without anterior knee pain. Causes are numerous and include ligamentous deficiency, muscular deficiency, anomlies of bony alignment and patellofemoral joint abnormalities. The 130 different procedures described to treat this condition reflect the multiple pathologies responsible. No single procedure has gained widespread acceptance. We present a surgical technique that attempts to correct as many of these deficiencies of the patellofemoral mechanism as possible.

Method: The procedure consists of a lateral release, a vastus medialis tendon advancement and a tibial tubercle osteotomy. The ‘Q’ angle is corrected by medialisation of the tubercle, patella alta is corrected by a distalisation technique and joint reaction forces through the patellofemoral joint are reduced by placing the tibial tuberosity in a more anterior position.

100 patients who have undergone this procedure have been identified. 81 percent initially presented with patella subluxation or dislocation. The remainder complained of anterior knee pain with evidence of abnormal patella tracking on examination. 52 percent of our patients had undergone at least one previous patellofemoral realignment procedure which had failed. 43 percent of the patients had generalised ligamentous laxity.

Results: The mean follow-up was 2.6 years from the index operation. 81 percent of the patients stated the operation had improved or abolished their symptoms. Generalised ligamentous laxity was present in the remaining 19 percent and seemed to correlate with a poor outcome. 66 percent of patients stated they were satisfied with the outcome of the surgery. Two patients developed recurrent subluxation after surgery and one of these has undergone a revision distal realignment procedure.

Using the functional category described by Crosby and Insall for patellofemoral symptoms 66 percent had a good-to-excellent outcome, 23 percent had a fair outcome and 10 percent of the patients stated they were worse following the procedure with increased anterior knee pain. 100 percent of these patients had grade 3 or 4 cartilage defects on retropatella surface. 57% of patients returned to sporting activity. 14% of the remainder had not returned to sporting activity because of persisting symptoms in the knee.

57 percent of patients had lost a mean 12.5 degrees of flexion of the knee at follow-up [range 5–30]. 5 percent of patients developed minor complications following surgery. No radiological deterioration was seen in any patients although coexistent patellofemoral osteoarthritis was seen in 25 percent of patients. There was no loss of fixation in any of the patients.

Discussion: A multifaceted approach to the complex problem of patellofemoral dysfunction appears to achieve satisfactory functional results in patients even when previous surgical realignment has failed. The procedure appears to be associated with low morbidity although a loss of flexion of the knee is to be expected. Generalised ligamentous laxity and cartilage defects on the retropatella surface appear to be associated with poor results and anterior knee pain in the absence of instability may be a cause for persisting symptoms.