A literature review of bone graft substitutes for spinal fusion was undertaken from peer reviewed journals to form a basis for guidelines on their clinical use. A PubMed search of peer reviewed journals between Jan 1960 and Dec 2009 for clinical trials of bone graft substitutes in spinal fusion was performed. Emphasis was placed on RCTs. Small and duplicated RCTs were excluded. If no RCTs were available the next best clinical evidence was assessed. Data were extracted for fusion rates and complications. Of 929 potential spinal fusion studies, 7 RCTs met the inclusion criteria for BMP-2, 3 for BMP-7, 2 for Tricalcium Phosphate and 1 for Tricalcium Phosphate/Hydroxyapatite (TCP/HA). No clinical RCTs were found for Demineralised Bone Matrix (DBM), Calcium Sulphate or Calcium Silicate. There is strong evidence that BMP-2 with TCP/HA achieves similar or higher spinal fusion rates than autograft alone. BMP-7 achieved similar results to autograft. 3 RCTs support the use of TCP or TCP/HA and autograft as a graft extender with similar results to autograft alone. The best clinical evidence to support the use of DBMs are case control studies. The osteoinductive potential of DBM appears to be very low however. There are no clinical studies to support the use of Calcium Silicate. The current literature supports the use of BMP-2 with HA/TCP as a graft substitute. TCP or HA/TCP with Autograft is supported as a graft extender. There is not enough clinical evidence to support other bone graft substitutes. This study did not require ethics approval and no financial support was received

Objectives

We performed a systematic review of the literature to determine the safety and efficacy of bone morphogenetic protein (BMP) compared with bone graft when used specifically for revision spinal fusion surgery secondary to pseudarthrosis.

We reviewed 15 consecutive patients, 11 women and four men, with a mean age of 48.7 years (37.3 to 62.6), who between July 2004 and August 2007 had undergone percutaneous sacroiliac fusion using hollow modular anchorage screws filled with demineralised bone matrix.

Each patient was carefully assessed to exclude other conditions and underwent pre-operative CT and MR scans. The diagnosis of symptomatic sacroiliac disease was confirmed by an injection of local anaesthetic and steroid under image intensifier control.

The short form-36 questionnaire and Majeed’s scoring system were used for pre- and post-operative functional evaluation. Post-operative radiological evaluation was performed using plain radiographs.

Intra-operative blood loss was minimal and there were no post-operative clinical or radiological complications. The mean follow-up was for 17 months (9 to 39). The mean short form-36 scores improved from 37 (23 to 51) to 80 (67 to 92) for physical function and from 53 (34 to 73) to 86 (70 to 98) for general health (p = 0.037). The mean Majeed’s score improved from 37 (18 to 54) pre-operatively to 79 (63 to 96) post-operatively (p = 0.014). There were 13 good to excellent results. The remaining two patients improved in short form-36 from a mean of 29 (26 to 35) to 48 (44 to 52). Their persistent pain was probably due to concurrent lumbar pathology.

We conclude that percutaneous hollow modular anchorage screws are a satisfactory method of achieving sacroiliac fusion.