416 - Fusion Rates and Predictive Factors for Pseudarthrosis Following Minim...

General Session: MIS-2

Presented by: S. Overley - View Audio/Video Presentation (Members Only)


S. Overley(1), S. McAnany(2), M. Anwar(1), A. Lovy(1), J. Guzman(1), D. Alicea(1), S. Qureshi(1)

(1) The Mount Sinai Hospital, Orthopaedic surgery, New York, NY, United States
(2) Emory University, Orthopaedic Surgery, Atlanta, GA, United States


Background Context: The advent of Bone Morphogenetic Protein (BMP) and live mesenchymal stem cell-containing Cellular Bone Matrices (CBM) as fusion adjuncts in MI-TLIF has resulted in new clinical data with a wide range of reported fusion rates.When employed in combination with an interbody cage in MI-TLIF procedures, the data shows fusion rates approach >90%;however, these studies rely on surgeon designation of fusion. No study to date has directly compared these two fusion substrates while utilizing a blinded third-party board certified neuroradiologist to assign fusion status.

Purpose: Our goal was to determine what effect, if any, graft adjuncts have on fusion rates, complications, and patient-reported outcome measures (PROM).Additionally, we set forth to identify any predictive factors for non-union via a regression analysis model.MI-TLIF is becoming an increasingly popular vehicle for lumbar fusion procedures.

Study Design and Setting: A retrospective analysis of a consecutive cohort of patients treated by a single surgeon, who underwent Minimally Invasive Transforaminal Lumbar Interbody Fusion (MI-TLIF) at an urban academic medical center.

Patient Sample: 52 consecutive, non-randomized patients undergoing MI-TLIF.

Outcome Measures: Fusion status, Oswestry Disability Index (ODI), Short-Form 12 (SF-12), and Euroqol 5-D (EQ-5D).

Methods: A retrospective review of a single surgeon's patients with symptomatic lumbar stenosis with spondylolisthesis who underwent a MI-TLIF with either BMP or CBM graft within a PEEK interbody cage was performed.Patient-reported outcome measures were recorded. CT scans were obtained at 12 month follow-up and were assessed by a board certified neuroradiologist.A successful fusion was based on bony bridging through the cage or outside of the cage in the anterior-posterior and lateral zones. T tests and Chi-square analysis were performed.A multi-variate regression analysis was performed to identify those patient factors that are predictive of non-union after MI-TLIF. Pearson r correlation coefficients were calculated to determine the impact of fusion status on clinical outcomes.

Results: 59 fusion levels in 52 patients were reviewed.37 patients received CBM;15 patients received rhBMP-2.The patients receiving BMP were older on average (63.9 vs 55.3, p=0.02).There were no statistical differences between the two groups with respect to BMI, insurance status, number of operative levels, smoking status, hypertension, or diabetes.The overall fusion rate was 50% in the Trinity group (20/40 levels) and 58% in the BMP group (11/19);the rate of fusion was not significantly different between the two groups (p=0.57).Multi-variate analysis showed that only multilevel fusion was predictive of nonunion.Overall, fusion status was not found to have a significant correlation on the patient reported outcomes.

Conclusions: The fusion rates demonstrated in this study are lower than what has been reported in the literature for MI-TLIF procedures.However, previous studies introduce considerable bias by utilizing the operating surgeon to assign fusion status rather than a blinded third party radiologist.Although the pseudarthrosis rate in this study is high, it has no bearing on outcome measures.Additionally, there was no difference in fusion rates between the two adjuncts.Only multilevel procedures were found to be predictive of non-union.Apart from fusion status, all other reported variables including re-operation and complication rates, ODI, SF-12, and EQ-5D were consistent with historical values.We hypothesize that MI-TLIF fusion rates reported in the literature may be inflated secondary to observer bias, though larger longitudinal blinded studies are needed to validate this claim.