Background. Extendable proximal femoral replacements(PFR) are used in children with bone tumours in proximity to the proximal femoral physis, previously treated by hip disarticulation. Long-axis growth is preserved, allowing limb salvage. Since 1986, survival outcomes after limb salvage and amputation have been known to be equal. Method. Retrospective review of all patients <16years undergoing extendable PFR at Royal National Orthopaedic Hospital (UK) between 04/1996 and 01/2006, recording complications, failures, procedures undertaken and patient outcomes. Results. 8 patients (mean age 8.9±3 years) underwent extendable PFR for Ewing's Sarcoma(5), Osteosarcoma(1), Chondrosarcoma(1) and rhabdomyosarcoma(1). 2 primary PFRs failed (infection of unknown source & local recurrence, both at 26months); 2 required revision for full extension (1 became infected at revision, requiring 2 stage revision). 3 patients had the original prosthesis in situ at last follow-up (mean 7.2;range 3–10.5years). 1 patient had no implant complications, but died (neutropaenic sepsis) 63 days after implant insertion. 2 were treated for recurrence but disease free at last review. 5 were continuously disease free. 5 patients were lengthened a mean 3.7cm; 2 were not lengthened.1 had incomplete data. 5 patients suffered subluxation/dislocation (mean 15.6months), 3 recurrently. Each underwent a mean 1.6 open & 1.4 closed procedures for the displaced joint. 3 patients had 4 open reductions and acetabuloplasties and 2 patients were converted to THR, with 3 major complications: 2 sciatic nerve palsies and 1 (THR) infection. The 5th patient was due for acetabuloplasty but had hip disarticulation for recurrence. Acetabular erosion occurred in 3; 2 were revised to THR (3.5 & 6.8years). 3 patients suffered peri-prosthetic supracondylar fracture (treated conservatively). 5 patients were revised to THR (mean 5.9years): 2 for dislocation, 2 for acetabular erosion & 1 for infection. 1 underwent amputation and another died. Only 1 surviving implant was not converted to THR: this patient had progressive acetabular erosion at 10.5 years & will eventually require THR. The amputee had poor hip function prior to disarticulation but went on to become an international Paralympic sportsman and had very good function 11.4 years post-disarticulation. 3 patients had fixed hip adduction deformity. 1 was isolated and treated with adductor tenotomy, whilst 2 were associated with knee flexion deformity (one required in-patient physiotherapy; the other prosthetic shortening). Conclusions. Extendable PFR permits limb salvage with psychological & functional benefits, but complications are common and some are specific to PFR. Surgery for these may result in further complications. Patients should be warned of the high conversion rate to THR. All the above should be borne in mind when selecting patients. As illustrated above, functional outcome is sometimes better with amputation

Background. Extendable partial femoral replacements (EPFR) permit limb salvage in children with bone tumours in proximity to the physis. Older designs were extended through large incisions or minimally invasive surgery. Modern EPFR are lengthened non-invasively. Lengthening improves functional score (Futani, 2006) but has been associated with complications including infection (Jeys, 2005). This study is the first to look specifically at the relationship between EPFR lengthening and complications. Method. Retrospective review of 51 paediatric (<16 years) oncology patients undergoing primary (1 °) EPFR (minimally/noninvasive) between 06/1994 and 01/2006. Exclusions: 1 patient with 5cm extension without medical intervention and 5 patients with incomplete data. Results. There was a negative relationship between age at 1 ° operation and EPFR lengthening (p = 0.03). Mean lengthening for all (including revision) EPFR was 3cm (range 0–13.05). 28 (64%) patients were lengthened a mean 4.8cm at a mean 7 procedures each. 18 were not lengthened, of which 12 were deceased at last followup and 1 had undergone amputation. Lengthening of secondary prostheses contributed a mean 3.7cm to limb length. 2 tertiary implants were lengthened noninvasively by 1.2cm (3 lengthenings) and 1.275cm (5 lengthenings). 16 of 21 failed 1 ° femoral components had been lengthened. The 21 implants underwent a mean 4.2 lengthenings in comparison to mean 1.3 lengthenings in the 28 that did not fail. Number of 1 ° prosthetic lengthenings was positively related to risk of component failure (p = 0.035) and total number of complications (p = 0.049). Complication rate was also related to total lengthening of 1 ° prosthesis (in cm) (p = 0.004). There was a statistically insignificant (p = 0.059) trend towards increased infection rate with higher number of minimally invasive lengthening procedures. Complications were commoner in those undergoing lengthening (mean 3 vs 1.75 in non-lengthened prostheses). Number of lengthenings correlated with number of complications (p = 0.001). Total lengthening of distal EPFR correlated with incidence of knee fixed flexion deformity (FFD)(p = 0.034). Median lengthening was 3cm in those with FFD and 0cm in those without (p = 0.019). Open lengthening procedures caused 6 complications: 2 deep & 3 superficial infections and recurrent subluxation requiring acetabuloplasty in a proximal replacement. Summary/Conclusions. The inverse relationship between age at 1 ° operation and lengthening is expected if the aim of offsetting limb length discrepancy during growth is being achieved. Lengthening is associated with increased complications (particularly infection and FFD) and femoral component failure, suggesting that lengthening procedures are implicated in the cause of some complications. 5 infective complications after lengthening were theoretically avoidable if noninvasive growers were used