Aims. Extendible endoprostheses have been available for more than 30 years and have become more sophisticated with time. The latest generation is ‘non-invasive’ and can be lengthened with an external magnetic force. Early results have shown a worryingly high rate of complications such as infection. This study investigates the incidence of complications and the need for further surgery in a cohort of patients with a non-invasive growing endoprosthesis. Patients and Methods. Between 2003 and June 2014, 50 children (51 prostheses) had a non-invasive growing prosthesis implanted for a primary bone sarcoma. The minimum follow-up was 24 months for those who survived. Their mean age was 10.4 years (6 to 14). The incidence of complications and further surgery was documented. Results. The mean follow-up was 64 months (20 to 145). The overall survivorship of the patients was 84% at three years and 70% at five years. Revision-free survival was 81.7% at three years and 61.6% at five years with competing risk analysis. Deep infection occurred in 19.6% of implants at a mean of 12.5 months (0 to 55). Other complications were a failure of the lengthening mechanism in five prostheses (9.8%) and breakage of the implant in two (3.9%). Overall, there were 53 additional operations (0 to 5 per patient). A total of seven patients (14%) underwent amputation, three for local recurrence and four for infection. Their mean limb length discrepancy was 4.3 mm (0 to 25) and mean Musculoskeletal Tumor Society Score functional score was 26.5 (18 to 30) at the final follow-up. Conclusions. When compared with previously published early results, this mid-term series has shown continued good functional outcomes and compensation for leg-length discrepancy. Infection is still the most common complication: post-operative wound healing problems, central line infection and proximal tibial location are the main risk factors. Cite this article: Bone Joint J 2016;98-B:1697–1703

Type BI rotationplasty is currently indicated for children with tumours of the proximal femur whereas type BIIIa rotationplasty is reserved for those in which the entire femur has to be removed. Our aim was to compare these two types of rotationplasty and determine whether the knee should be preserved in children with tumours of the proximal femur. We compared the post-operative complications, oncological outcome, range of movement, Enneking score and radiographs of six children, who had undergone type BI rotationplasty with those of 12 who had undergone type BIIIa rotationplasty.

Patients with type BI rotationplasty had a mean Enneking score of 21.6 compared with 24.4 in those with type BIIIa rotationplasty, and worse mean results in all of the parameters investigated. We conclude that type BI rotationplasty has a worse functional outcome and more complications than type BIIIa rotationplasty in children under the age of ten years.