Introduction Morbidity associated with wound complications may translate into disability and quality-of-life disadvantages for patients treated with radiotherapy (RT) for soft tissue sarcoma (STS) of the extremities. Functional outcome and health status of extremity STS patients randomized in a phase III trial comparing pre-operative versus post-operative RT is described.

Methods One hundred and ninety patients with extremity STS were randomized after stratification by tumor size dichotomized at 10 cm. Function and quality of life were measured by the Musculoskeletal Tumor Society Rating Scale (MSTS), the Toronto Extremity Salvage Score (TESS), and the Short Form-36 (SF-36) at randomization, six weeks, and three, six, 12, and 24 months after surgery. One hundred and eighty-five patients had function data.

Results Patients treated with post-operative RT had better function with higher MSTS (25.8 v 21.3, P < .01), TESS (69.8 v 60.6, P =.01), and SF-36 bodily pain (67.7 v 58.5, P =.03) scores at six weeks after surgery. There were no differences at later time points. Scores on the physical function, role-physical, and general health sub-scales of the SF-36 were significantly lower than Canadian normative data at all time points. After treatment arm was controlled for, MSTS change scores were predicted by a lower-extremity tumor, a large resection specimen, and motor nerve sacrifice; TESS change scores were predicted by lower-extremity tumor and prior incomplete excision. When wound complication was included in the model, patients with complications had lower MSTS and TESS scores in the first two years after treatment.

Conclusions The timing of RT has minimal impact on the function of STS patients in the first year after surgery. Tumor characteristics and wound complications have a detrimental effect on patient function.

Introduciton In our experience, amputation is rarely indicated in osteosarcoma. Amputation is more frequently required in soft tissue sarcoma for the following reasons: 1) recurrent tumour in previously radiated field; 2) composite tissue involvement of soft tissue, bone, vessels and nerves; 3) size of lesion. We have measured functional outcome in bone and soft tissue sarcoma using a combination of clinimetric measures describing impairment and patient determined measures assessing disability (1,2).

Methods In a matched case-control study (3), 12 patients with amputation were matched with 24 patients treated by limb-sparing surgery on the following variables: age, gender, length of follow-up, bone versus soft-tissue tumor, anatomic site, and treatment with adjuvant chemotherapy. End points included the Toronto Extremity Salvage Score (TESS), a measure of physical disability; the Shortform-36 (SF-36), a generic health status measure; and the Reintegration to Normal Living (RNL), a measure of handicap.

Results Mean TESS score for the patients with amputations was 74.5 versus 85.1 for the limb-sparing patients. (p = .15). Only the physical function subscale of the SF-36 showed statistically significant differences, with means of 45 and 71.1 for the amputation versus limb-sparing groups, respectively (p = .03). The RNL for the amputation group was 84.4 versus 97 for the limb-sparing group (p = .05). Seven of the 12 patients with amputations experienced ongoing difficulty with the soft tissues overlying their stumps. There was a trend toward increased disability for those in the amputation group versus those in the limb-sparing group, with the amputation group showing significantly higher levels of handicap.

Conclusions These data suggest that the differences in disability between amputation and limb-sparing patients are smaller than anticipated. The differences may be more notable in measuring handicap.

Introduction Localised Langerhans-cell histiocytosis of bone (eosinophilic granuloma) is a benign tumour-like condition with a variable clinical course. Different forms of treatment have been reported to give satisfactory results. However, previous series all contain patients with a wide age range. Our aim was to investigate the effect of skeletal maturity on the rate of recurrence of isolated eosinophilic granuloma of bone excluding those arising in the spine.

Methods We followed-up 32 patients with an isolated eosinophilic granuloma for a mean of five years; 17 were skeletally immature.

Results No recurrences were noted in the skeletally immature group even after biopsy alone. By contrast, four of 13 skeletally mature patients had a recurrence and required further surgery.

Conclusions This suggests that eosinophilic granuloma has a low rate of recurrence in skeletally immature patients.

Aims: To review the results of reconstruction of pseudoarthrosis and/or significant varus with retroversion of proximal femur in congenital longitudinal lower limb deficiencies. Methods: 23 of 95 patients had proximal femoral reconstruction. 7 had pseudoarthrosis of the neck of femur and the remaining had significant coxa vara with retroversion of femur. 3 patients with pseudoarthroses were treated with valgus derotation osteotomy and cancellous bone grafting, 2 with fibular strut grafts, 1 King’s procedure and 1 with excision of fibrous tissue and valgus derotation osteotomy. Remaining patients underwent valgus derotation osteotomies. A variety of internal fixation devices and external fixators were used. Results: All patients with pseudoarthroses underwent multiple procedures to achieve union. 3 with cancellous bone grafting underwent repeat osteotomies to correct residual varus and 2 had grafting repeated twice. 2 patients achieved union after fibular strut grafting. One patient, who underwent excision of pseudoarthrosis, achieved union but had to undergo further valgus osteotomy. The remaining 17 patients with coxa vara and retroversion of femur also had valgus osteotomies repeated more than once (average 2.3) for recurrence of varus deformity. There were significant numbers of implant failures. Average initial neck-shaft angle of 72 degrees improved to 115 degrees after reconstruction. Conclusion: Achieving union of pseudoarthrosis and early axis correction using valgus derotation osteotomy with a view to later lengthening is important in limb reconstruction. Recurrence may require repeated osteotomies and pseudoarthrosis may need more aggressive surgery to achieve union. Muscle slides and soft tissue releases decrease the stress on implant and maintain correction.

The treatment of acetabular dysplasia in adolescents (age> 12) is difþcult and various complex pelvic osteotomies have been described. The aim of surgery being improvement in pain and to delay the onset of secondary osteoarthrosis. Methods: We present our experience of using the Tonnis triple pelvic osteotomy for treatment of acetabular dysplasia in the adolescent and adult age groups (range 13–27 years). This retrospective analysis includes 25 patients operated on over a nine year period (1991–2000) with an average followup of four years (range 2–8 years). More than 50% of the patients had had a previous open reduction or femoral osteotomy for CDH. Results: Radiographic analysis (pre & post op) included CE angle of Wiberg, Sharp-Ullmann index and the acetabular angle of the weight bearing zone. All parameters showed an improvement in the post operative analysis with an improvement in pain and range of movement in all patients. Discussion: The Tonnis triple pelvic osteotomy has the advantage of allowing the operator a direct þeld of view at all times and achieving a great deal of lateral rotation and medial displacement of the acetabulum due to the proximity of the osteotomy to the acetabulum. The ischial ramus and its ligaments to the sacrum are left intact, leading to greater stability of the pelvis and spine. Conclusion: Though technically difþcult and needing a long learning curve, it does improve acetabular alignment and symptoms in the early postoperative years. However long term studies are required to document its effect on the rate of secondary osteoarthrosis.

Aims: To review the results of limb lengthening and deformity correction in fibular hemimelia. Methods: Fifty-five patients with fibular hemimelia underwent limb reconstruction at Sheffield Children’s Hospital. According to Achterman and Kalamchi classification, twenty-six were classified as Type IA, six as Type IB and twenty-three as Type II fibular hemimelia. All patients had at least some shortening of ipsilateral femur but forty-nine had significant femoral deficiency. Lengthening of tibia and in significant cases femur was done using either De Bastiani, Vilarrubias or Ilizarov methods. Ankle valgus and heel valgus were corrected through osteotomies either in the supramalleolar region or heel. Equinus was corrected by lengthening of tendoachelis with posterior soft tissue release and in severe cases using Ilizarov technique. Results: The average length gained was 4.2 cm (range 1 to 8) and the mean percentage of length increase was 15.82 (range 4.2 to 32.4). Mean bone healing index was 54.23 days/cm. Significant complications included knee subluxation, transient common paroneal nerve palsy, and recurrence of equinus and valgus deformity of foot. Overall alignment and ambulation improved in all patients. Knee stiffness due to cruciate deficient subluxations needed prolonged rehabilitation. Presence of 3-ray foot gives a better functional result and cosmetic acceptance by patients. The Ilizarov frame has the advantage to cross joints and lengthen at the metaphysis. Conclusion: Limb reconstruction in fibular hemimelia using limb lengthening and deformity correction techniques improve functional status of involved lower limb.

Aims: Acute lateral dislocation of the patella has been associated with disruption of the medial restraints of the patella and following non-operative management, a redislocation rate of up to 44%. Methods: Ten patients who presented to the Accident and Emergency dept. following acute patella dislocation had an ultrasound scan (USS) performed by an experienced musculoskel-etal radiologist. Each patient had an arthroscopy and washout of the knee and repair of ruptured structures. The ultrasound reports have now been compared to the surgical findings to determine the effectiveness of this investigation. Results: The ultrasound scans identified deficiencies in the ligamentous attachments to the medial border of the patella in eight patients and these were confirmed at operation in the same eight. The USS diagnosis of haematoma or torn fibres in the vastus medialis obliquus (VMO) (5 patients) corresponded with tearing of this muscle at operation; however the degree of muscle injury was underestimated in two. The USS finding of free fluid around the medial collateral ligament (MCL) at the adductor tubercle in three patients was associated with the operative finding of disruption of the femoral origin of the medial patellofemoral ligament (MPFL). Haematomata detected on USS along the adductor longus in two patients proved to have disruption of the VMO attachment. Conclusions: We recommend the use of ultrasound for assessment of all patella dislocations to accurately locate tears of the retinaculum and help clinicians to understand the severity of injury to the soft tissue restraints of the patella.

Aims: To review the hip subluxations or dislocations occurring during femoral lengthening in patients with congenital longitudinal lower limb deficiencies. Methods: Sixty-three patients with congenital longitudinal lower limb deficiencies underwent femoral lengthening using either De Bastiani, Villarubias or Ilizarov technique. Acetabular index, medial joint space, CE angle of Wiberg, acetabular angle of Sharp and neck-shaft angle were measured on anteroposterior radiographs of hip before, during and after lengthening. Acetabulum was considered dysplastic when the Sharp angle was more than 45 degrees. Hip was considered to be subluxed when the medial joint space increased during lengthening. Results: During femoral lengthening, eleven hips subluxed as measured by the increase in medial joint space and one hip dislocated. All these hips had a pre-operative acetabular index more than 25 degrees, CE angle less than 20 degrees and Sharp angle more than 45 degrees. The average neck-shaft angle was 75 degrees. Following subluxation, lengthening was stopped and the hips were reduced in hip spica after adductor and sartorius tenotomies. In one patient femoral shortening and acetabuloplasty had to be done to reduce the sub-luxation. No case of avascular necrosis or chondrolysis was noted. Conclusions: Hip subluxation during femoral lengthening of congenital longitudinal lower limb deficiencies tends to occur when there is associated ace-tabular dyplasia and femoral coxa vara. Careful preoperative assessment, if need be hip reconstruction prior to lengthening and close monitoring during lengthening is recommended.

We performed limb lengthening and correction of deformity of nine long bones of the lower limb in six children (mean age, 14.7 years) with osteogenesis imperfecta (OI). All had femoral lengthening and three also had ipsilateral tibial lengthening. Angular deformities were corrected simultaneously. Five limb segments were treated using a monolateral external fixator and four with the Ilizarov frame. In three children, lengthening was done over previously inserted femoral intramedullary rods.

The mean lengthening achieved was 6.26 cm (mean healing index, 33.25 days/cm). Significant complications included one deep infection, one fracture of the femur and one anterior angulation deformity of the tibia. The abnormal bone of OI tolerated the external fixators throughout the period of lengthening without any episodes of migration of wires or pins through the soft bone. The regenerate bone formed within the time which is normally expected in limb-lengthening procedures performed for other conditions.

We conclude that despite the abnormal bone characteristics, distraction osteogenesis to correct limb-length discrepancy and angular deformity can be performed safely in children with OI.

Aim: The aim of the study was to evaluate the outcome of open surgery for DDH in a tertiary centre.

Method: Primary and tertiary referrals between 1983 and 1996 were followed up for 8.2 years (4.1 to 16.8 years). 60 hips in 47 patients had open reduction for DDH. 39 patients were females. 13 hips had bilateral involvement. Age at operation was 2.1 years (0.1 to 5.5 years). 32 patients were referred from elsewhere for surgery after closed or open primary treatment.

Results: Clinical and radiological assessment was done on all the patients. The outcome was evaluated using the McKay system and Severin grade for subluxation. 12 hips had Type II, III or IV or unclassifiable AVN according to the Kalamchi and MacEwen classification (20%). The other results are tabulated as follows:

Complications other than AVN were re-subluxation (3), redislocation(4), fractures (1), ankylosis, LLD(4), infection (2). There was only 1 (5.9%) complication in primary referral group and 13 (30.2%) in tertiary referral group (p=0.050). Tonnis grade of subluxation, presence of ossific nucleus, tear drop shape, Mose’s grading, CE angle of Wiberg were documented but were not found to significantly affect the outcome.

Conclusion: Multiple operation increases the risk of proximal femoral growth disturbance and affects the outcome. No statistical difference was noted in the AVN rate between the primary referral group and the tertiary referral group. However, there were significantly more complications in the tertiary referral group and there maybe a need for early tertiary referral.

Aim: To determine the ultra-structural morphology of bone in CLLLD.

Materials and Methods: Bone biopsies were taken from 8 patients with CLLLD undergoing surgery for limb reconstruction procedures. Specimens were fixed and processed for Electron microscopy using standard processing protocol. Ultra-thin araldite sections were stained with uranyl acetate and lead citrate and viewed in Philips CM12 electron microscope. Muscle biopsies were also undertaken.

Results: There were 5 boys and 3 girls of a mean age of 8.8 years (range 3 to 14 years). 6 had predominantly femoral deficiencies and 2 had combined femoral and fibular deficiencies. All specimens showed abnormal collagen fibril morphology. They showed variable diameter with irregular outlines in transversely section bundles and appeared unravelled in longitudinally sectioned bundles. The osteoid was disorganised in 4 of the 8 specimens and osteoblasts showed necrotic changes in 5 out of 8 specimens. Patchy mineralisation and increased proteoglycan distribution was observed in 3 out of 8 specimens.

Conclusion: Bone in CLLLD shows ultra-structural changes in collagen and osteoblasts that may account for the retarded bone growth and poor regenerate formation that occurs during limb lengthening in these patients.

Upper limbs are commonly involved in Arthrogyposis Multiplex Congenita. They may be involved in isolation or in combination with lower limbs. There are two patterns of involvement in upper limbs. The most common (type I) pattern presents with adduction and internal rotation at the shoulder, extension at the elbow, pro-nation of the forearm and flexion deformity of the wrist, indicating involvement of the C5 and C6 segments. These deformities can be quite disabling and may require surgery to help improve function. We present our long-term results with pectoralis major transfer procedure (as modified by senior author MJB) to restore elbow flexion in seven patients (ten procedures).

Results: Early results in all our patients were quite encouraging. Six patients retained useful power in transferred pectoralis major muscle and maintained the arc of flexion, which was attained following tricepsplasty. However, as children were followed up a gradually increasing flexion deformity and decreasing flexion arc were observed in eight elbows. The onset and progression of flexion deformity was gradual and progressive. The flexion deformity reached ninety degrees or more in all cases.

Conclusions: Results of pectoralis major transfer to treat extension contracture of the elbow in arthrogryposis deteriorate with time due to development of recalcitrant flexion deformity of the elbow. Presently we recommend this procedure on one side only in cases of bilateral involvement because if one procedure is carried out it would be possible for this hand to get to the mouth for feeding and the other unoperated side would be able to look after the perineal hygiene

Children with osteogenesis imperfecta(OI) have multiple long bone fractures with subsequent deformities. The mainstay of treatment is correction with multiple osteotomies and intramedullary fixation. The Shefffield intramedullary telescoping rod system has been successful in the treament of long bone fractures and deformities (Wilkinson et al ,JBJS-B,1998) Bisphosphonates (Pamidronate -1- 1.5mg/kg/day)have been used as adjuvant therapy in the treatment of OI since the last five years. The perceived benefits include reduction in fracture frequency, improvement in bone density and a general feeling of well being.

We present our experience of five cases of OI who developed infections around thier Sheffield telescoping rods while on Pamidronate therapy. There was only one case of sepsis over a ten year period(over eighty patients)in a previously reported series from our centre.

The time interval between the start of Pamidronate therapy and the diagnosis of infection varied between 12–36 months ie. between 4–12 cycles of Pamidronate (parenteral administration over a three day period at three month intervals). All patients had their intramedullary rods in situ from anywhere between 2–7 years. The infections were low grade with a 2–3 month period of dull ache prior to actual presentation. Intrestigly though all patients had multiple rods in situ, only one of their femoral rods was affected and they did not have any other infective focus at the time of diagnosis. Three patients presented with thigh abcesses while the other two presented with ipsilateral knee pain and effusion. All had raised inflammatory markers, radiological signs of sepsis with Staph Aureus the commonest infecting organism.

Those cases presenting with abcesses were treated by drainage and rod removal, however only antibiotics were sufficient in the rest. The relationship between Pamidronate therapy and these infections is not absolutely clear and has not been reported previously. The possible links are discussed and a high degree of suspicion is recommended for those cases of OI on bisphosphonate

Objective: To review the hip subluxations or dislocations occurring during femoral lengthening in patients with congenital longitudinal lower limb deficiencies.

Methods: Sixty-three patients with congenital longitudinal lower limb deficiencies underwent femoral lengthening using either De Bastiani, Villarubias or Ilizarov technique. Acetabular index, medial joint space, CE angle of Wiberg, acetabular angle of Sharp and neck-shaft angle were measured on anteroposterior radiographs of hip before, during and after lengthening. The Acetabulum was considered dysplastic when the Sharp angle was more than 45 degrees. Hip was considered to be subluxed when the medial joint space increased during lengthening.

Results: During femoral lengthening, eleven hips sub-luxed as measured by the increase in medial joint space and one hip dislocated. All these hips had a preoperative acetabular index more than 25 degrees, CE angle less than 20 degrees and Sharp angle more than 45 degrees. The average neck-shaft angle was 75 degrees. Following subluxation, lengthening was stopped and the hips were reduced in hip spica after adductor and sartorius tenotomies. In one patient femoral shortening and acetabulo-plasty had to be done to reduce the subluxation. No case of avascular necrosis was noted.

Conclusion: Hip subluxation during femoral lengthening of congenital longitudinal lower limb deficiencies tends to occur when the acetabular index is more than 25 degrees, Sharp angle is more than 45 degrees, CE angle is less than 20 degrees and when there is associated femoral coxa vara. Careful preoperative assessment is required, and if need be hip reconstruction prior to lengthening. Close monitoring during lengthening is recommended.

Study background: The management of patients presenting in primary care with a “new episode of back pain” using the RCGP guidelines by a nurse practitioner (NP) compared to a control group given ‘usual care’ by the GP.

Methods and Results: The evaluation includes sequential monitoring of LBOS, audit of documentation for both groups of patients to assess application of guidelines and patient recall of key information.

Conclusion: The NHS Plan (2000) called for national standards for treating all major conditions in “Shifting the Balance of Power- the next steps” (2001) a clear criterion is that “service outcomes which provide better and better validated information” will form part of the performance rating. These two documents highlight the need for more user and client involvement in service development but this needs to be carefully monitored and linked to effective evidence.

This study demonstrates that the NP documentation follows the guidelines identified by the RCGP, conversely it was not possible to assess from the GP documentation if all the steps had been followed. The mean average LBOS in the NP patients was slightly higher than those in the GP group, was this because these patients were having guideline applied care as opposed to “usual care”?

Evaluation of the patient recall of information shows the NP sent five patients for X-ray even though this did not occur and is not recommended in guidelines. Conversely twenty-three patients can remember being given the “Back Book” by the GP but this was only documented in three cases.

We believe that patient recall demonstrates an ineffective way to measure outcome and funding allocation for back pain management and needs to more accurately reflect the evidence.

Increased incidence of complications has been reported when lengthening limbs with underlying bone disorders such as dysplasias and metabolic bone diseases. There is a paucity of literature on limb lengthening in Osteogenesis Imperfecta (OI), probably due to the concern that the bone containing abnormal collagen may not tolerate the external fixators for a long term and there may not be adequate regenerate formation from this abnormal bone.

We performed limb lengthening and deformity correction of nine lower limb long bones in six children with OI. Four children were type I and two were type IV OI as per Sillence classification. The mean age was 14.7 years. All six children had lengthening for femoral shortening and three of them also had lengthening for tibial shortening on the same side. Angular deformities were corrected during lengthening. Five limb segments were treated using a monolateral external fixator and four limb segments were treated using an Ilizarov external fixator. In three children, previously inserted femoral intramedullary nails were left in situ during the course of femoral lengthening. The average lengthening achieved was 6.26 cm. Limb length discrepancies were corrected to within 1.5 cm of the length of the contralateral limb in five children. In one child with fixed pelvic obliquity and spinal scoliosis, functional leg length was achieved. The mean healing index was 33.25 days/cm of lengthening. Among the complications significant ones included, one deep infection, one fracture through the midshaft of the femur, and development of anterior angulation deformity after the removal of the fixator in one tibia. Abnormal bone of OI tolerated the external fixator throughout the period of lengthening without any incidence of migration of wires and screws through the soft bone when distraction forces were applied. The regenerate bone formed within the time that is normally expected in limb lengthening procedures performed for other conditions. We conclude that despite abnormal bone characteristics, limb reconstruction to correct limb length discrepancy and angular deformity can be done safely in children with OI.

Acute lateral dislocation of the patella has been associated with disruption of the medial restraints of the patella. Following non-operative management there is a redislocation rate of up to 44%. This is an observational study testing whether sonography is a reliable method of assessing the medial retinaculum after acute dislocation of the patella.

Ten patients following acute patellar dislocation had an ultrasound scan (USS) performed by an experienced musculoskeletal radiologist. Each patient subsequently had an examination under anaesthetic, arthroscopy, and repair of the ruptured structures. The ultrasound reports were compared to the surgical findings to determine the accuracy of this investigation.

USS located deficiencies in the ligamentous attachments to the medial border of the patella and the presence of avulsed bony fragments, all of which were confirmed at operation. The sonographic diagnosis of haematoma or torn fibres in the vastus medialis obliquus corresponded with our operative findings.

The most significant findings were the correlation of free fluid around the medial collateral ligament (MCL) with avulsion of the femoral attachment of the medial patellofemoral ligament (MPFL) and the presence of avulsed fragments of bone from the medial border of the patella.

Sonography, in cadaveric studies consistently identifies the retinacula and like MRI offers a distinctive constellation of findings that can be used in diagnosis and therefore play a significant role in directing surgical management of these patients. We have found Sonography to be readily available and accurate.

This report does not include surgical outcome since the follow up is short and incomplete. We do, however, feel that ultrasound shows the state of the soft tissue restraints of the patella following lateral dislocation.

113 consecutive patients with soft tissue sarcoma treated by excision and reconstructive flaps were studied to assess the risk of complications and to compare local tumour control with those in whom primary wound closure was possible.

Minimum follow-up was 24 months and mean age was 55 years (16–95). The sarcoma was located in the lower extremity in 83 and upper extremity 30 patients. Significant wound complications developed in 37 patients (33%). The most common complications were wound infections or partial necrosis occurring in 16% (18/113) and 13% (15/113) respectively. Complete flap necrosis requiring flap removal occurred in 6 patients (5%). Three patients (2.3%) required amputation as a result of complications. Significant risk factors for development of wound complications include location of tumour in the lower limb compared to upper limb (relative risk 2.3, p=0.02) and use of pre-operative radiotherapy compared to no or post-operative radiotherapy (relative risk 2.05, p=0.02). There was no difference in rates of complications in patients with free or pedicled flaps, tumours < or > 5cm, distal or proximal location of tumour.

The rates of negative excision margins (80%) and wound complications in patients who required reconstructive flaps were not different from that for the other patients treated at our centre who did not require reconstructive flaps.

The use of soft tissue reconstructive flaps did not reduce the risk of positive excision margins or the rates of wound complications. The risk of amputation secondary to flap complication or failure is low.

812 consecutive patients with soft tissue sarcoma of the extremity were studied to compare the characteristics and outcome of patients who had primary amputations and limb preserving surgery.

Patients with primary amputations were more likely to have metastases at presentation, high-grade tumours, larger tumours and were older.

The most frequent indications for primary amputation were tumour excision which would result in inadequate function and large extracompartmental tumours with composite tissue involvement including major vessels, nerves and bone.

The requirement for primary amputation was a poor prognostic factor independent of tumour grade, tumour size and patients’ age.

To determine if rates of local recurrence and metastasis differ in upper versus lower extremity sarcomas.

Prospectively collected data relating to patients undergoing limb-sparing surgery for extremity soft tissue sarcoma between January 1986 and April 1997 were analysed. Local recurrence-free and metastasis-free rates were calculated using the method of Kaplan and Meier. Univariate and multivariate analyses of potential predictive factors were evaluated with the log-rank test and the Cox proportional hazards model.

Of 480 eligible patients, 48 (10. 0%) had a local recurrence and 131 (27. 3%) developed metastases. Median follow-up of survivors was 4. 8 years (0. 1 to 12. 9). There were 139 upper and 341 lower extremity tumours. Upper extremity tumours were more often treated by unplanned excision before referral (89 vs 160, p< 0. 001) and were smaller (6. 0cm vs 9. 3cm, p< 0. 000). Lower extremity tumours were more often deep to or involving the investing fascia (280 vs. 97, p< 0. 003). The distribution of histological types differed in each extremity. Fewer upper extremity tumours were treated with adjuvant radiotherapy (98 vs. 289, p< 0. 000).

The 5-year local recurrence-free rate was 82% in the upper and 93% in the lower extremity (p< 0. 002). Local recurrence was predicted by surgical margin status (hazard ratio 3. 16, p< 0. 000) but not extremity (p=0. 127) or unplanned excision before referral (p=0. 868).

The 5-year metastasis-free rate was 82% in the upper and 69% in the lower extremity (p< 0. 013). Metastasis was predicted by high histological grade (hazard ratio 17. 28, p< 0. 000), tumour size in cm (hazard ratio 1. 05, p< 0. 001) and deep location (hazard ratio 1. 93, p< 0. 028) but not by extremity (p=0. 211).

Local recurrence is more frequent after treatment for upper compared with lower extremity sarcomas. Variation in the use of radiotherapy and differences in histological type may be contributory. Metastasis is more frequent after treatment for lower extremity sarcomas because tumours tend to be large and deep.