Devoted to diagnostic and interventional spine imaging and therapeutics


Complications in the use of rh BMP-2 (Bone Morphogenic Protein) in PEEK cages for interbody spinal fusions 2006

General Spine

Joseph G Craig, MB ChB, Non ASSR Member
Mark Jacobson, B Sc, Non ASSR Member
Chad Coe, B Sc, Non ASSR Member
Rahul Vaidya, MD, Non ASSR Member

Scientific Paper


To review a series of patients in whom rh BMP-2 had been used in PEEK cages for interbody fusion with the purpose of accelerating/enhancing new bone formation at the site of fusion

Methods & Materials

We retrospectively reviewed 59 consecutive patients who had 82 interbody fusion levels (32 cervical, 50 lumbar) using PEEK cages with rh BMP-2. X-rays were examined at post operative intervals (2 weeks, 6 weeks; 3, 6, 9, 12 months). Radiological assessment included new bone formation, endplate resorption and regrowth, migration of the PEEK cage and subsidence of the intervertebral disc space. CT evaluation was available in 10 lumbar spine cases.


Resorption of the endplates/adjacent vertebral body in the cervical fusions was a prominent finding between 6 weeks and 12 weeks post operatively and had a similar appearance to infective discitis. New bone formation across the disc space in the PEEK cage was observed 3 months to 9 months post operatively with solid fusion between 9 and 12 months. The lumbar spine was more difficult to assess radiographically but resorption was seen 3-6 months after surgery. New bone formation in the lumbar spine was seen later than the cervical spine being seen between 3 and 9 months. CT evaluation showed similar results.

In the cervical spine, PEEK cage migration was seen in 1 of 32 levels but no revision was required.

In the lumbar spine, migration was seen in 1 of 12 anterior lumbar interbody fusions and 9 of 36 Transforaminal Lumbar Interbody Fusion (TLIF) levels, 8 of which required revision surgery.

Subsidence of the cervical disc space was seen in 13 levels with the average subsidence 12.8%. In the lumbar spine, subsidence was seen at 11 levels and the average subsidence was 17.8%.


As well as producing new bone formation in the cervical and lumbar spine, BMP initially produces a marked resorptive response which is expressed radiographically as endplate resorption. The appearances may be identical to infective discitis/osteomyelitis

Because of this response, surgical cages which were initially placed with solid bone apposition are more likely to migrate. This may lead to neurologic sequelae requiring revision surgery in TLIFS.

Subsidence was greater in the lumbar spine than the cervical spine. The cause of this is uncertain but subsidence was not a problem clinically