The effects of local insulin application to lumbar spinal fusions in a rat model

Abstract

The rates of pseudoarthrosis after a single-level spinal fusion have been reported up to 35%, and the agents that increase the rate of fusion have an important role in decreasing pseudoarthrosis after spinal fusion. Previous studies have analyzed the effects of local insulin application to an autograft in a rat segmental defect model. Defects treated with a time-released insulin implant had significantly more new bone formation and greater quality of bone compared with controls based on histology and histomorphometry. A time-released insulin implant may have similar effects when applied in a lumbar spinal fusion model. This study analyzes the effects of a local time-released insulin implant applied to the fusion bed in a rat posterolateral lumbar spinal fusion model. Our hypothesis was twofold: first, a time-released insulin implant applied to the autograft bed in a rat posterolateral lumbar fusion will increase the rate of successful fusion and second, will alter the local environment of the fusion site by increasing the levels of local growth factors. Animal model (Institutional Animal Care and Use Committee approved) using 40 adult male Sprague-Dawley rats. Forty skeletally mature Sprague-Dawley rats weighing approximately 500 g each underwent posterolateral intertransverse lumbar fusions with iliac crest autograft from L4 to L5 using a Wiltse-type approach. After exposure of the transverse processes and high-speed burr decortication, a Linplant (Linshin Canada, Inc., ON, Canada) consisting of 95% microrecrystalized palmitic acid and 5% bovine insulin (experimental group) or a sham implant consisting of only palmitic acid (control group) was implanted on the fusion bed with iliac crest autograft. As per the manufacturer, the Linplant has a release rate of 2 U/day for a minimum of 40 days. The transverse processes and autograft beds of 10 animals from the experimental and 10 from the control group were harvested atDay 4 and analyzed for growth factors. The remaining 20 spines were harvested at 8 weeks andunderwent a radiographic examination, manual palpation, and microcomputed tomographic (micro-CT) examination. One of the 8-week control animals died on postoperative Day 1, likely due to anesthesia. In the groups sacrificed at Day 4, there was a significant increase in insulinlike growth factor-I (IGF-I) in the insulin treatment group compared with the controls (0.185 vs. 0.129; p=.001). No significant differences were demonstrated in the levels of transforming growth factor beta-1, platelet-derived growth factor-AB, and vascular endothelial growth factor between the groups (p=.461, .452, and .767 respectively). Based on the radiographs, 1 of 9 controls had a solid bilateral fusion mass, 2 of 9 had unilateral fusion mass, 3 of 9 had small fusion mass bilaterally, and 3 of 9 had graft resorption. The treatment group had solid bilateral fusion mass in 6 of 10 and unilateral fusion mass in 4 of 10, whereas a small bilateral fusion mass and graft resorption were not observed. The difference between the groups was significant (p=.0067). Based on manual palpation, only 1 of 9 controls was considered fused, 4 of 9 were partially fused, and 4 of 9 were not fused. In the treatment group, there were 6 of 10 fusions, 3 of 10 partial fusions, and 1 of 10 were not fused. The difference between the groups was significant (p=.0084). Based on the micro-CT, the mean bone volume of the control group was 126.7 mm(3) and 203.8 mm(3) in the insulin treatment group. The difference between the groups was significant (p=.0007). This study demonstrates the potential role of a time-released insulin implant as a bone graft enhancer using a rat posterolateral intertransverse lumbar fusion model. The insulin-treatment group had significantly higher fusion rates based on the radiographs and manual palpation and had significantly higher levels of IGF-I and significantly more bone volume on micro-CT.