Introduction. We aimed to determine the extent of osseous integration of the hydroxyapatite collars of tumour endoprostheses implanted in our unit. Methods. We identified 57 patients who had massive endoprostheses implanted over the last six years and reviewed clinical records and x-rays. There were 20 proximal femoral, 23 distal femoral, 6 proximal tibial, 8 proximal humeral and 1 distal humeral replacements. Patients fell into three groups: 1.Primary bone tumours, 2. Metastatic bone tumours and 3. Non-tumour indications. Results. There were 32 patients in Group 1. Mean age 32(10-71) years. Mean follow up was 22 (12-60) months. Osseous integration of the collars was noted in 19 patients (60%). Eight patients had no bony ongrowth on the collars. Nine patients underwent revision, seven for aseptic loosening and one for infection. There were 23 patients in Group 2. Mean age 72(45-83) years. The primary tumour was lung in four and renal in three. One patient had bilateral surgery. Mean follow up was 6(4-9) months. One prosthesis required revision for aseptic loosening in spite of showing good growth over the collar. True osseous integration of the collar occurred only in one patient. There were two patients in Group 3: one with a distal femoral fracture non union who demonstrated good integration of the collar. The other had a revision for infected total hip replacement with no osseous integration. Discussion. Hydroxyapatite coated collars may help integration of massive endoprostheses, but ongrowth is not always seen. Integration is best after resection of primary bone tumours. Their role in metastatic tumour surgery is questionable