Costs and outcomes of recombinant human bone morphogenetic protein-2 (rhBMP-2) usage in single level posterior or transforaminal lumbar interbody fusion (TLIF) in a propensity score matched national, retrospective cohort.

BACKGROUND CONTEXT: Recombinant human bone morphogenetic protein-2 (rhBMP-2) is widely used in posterior or transforaminal lumbar interbody fusion (P/TLIF). While rhBMP-2 demonstrates comparable fusion rates to autologous iliac crest bone graft (AICBG), its high cost brings into question its long-term cost effectiveness. This question is particularly pertinent in single-level P/TLIF where AICBG achieves excellent fusion rates.

PURPOSE: To evaluate the costs associated with and clinical outcomes of rhBMP-2 versus non-rhBMP-2 usage in single-level P/TLIF using a propensity score-matched national sample.

STUDY DESIGN/

SETTING

Retrospective propensity score-matched cohort study using the Premier Healthcare claims database (2016-2022).

PATIENT SAMPLE: National sample of 9,457 patients who underwent single-level P/TLIF with rhBMP-2 utilization who were one-to-one matched with 9,457 P/TLIF patients without rhBMP-2 based on demographics, Elixhauser comorbidities, and surgical indications.

OUTCOME MEASURES: Primary economic outcomes were index encounter costs by category, and 2-year cumulative post-discharge costs. Primary clinical outcomes were 90-day readmission and 2-year revision fusion incidence. Secondary clinical outcomes were 2-year revision decompression incidence, pseudoarthrosis, seroma, infection, and implant complications.

METHODS

Single level P/TLIF procedures were identified via billing codes for the procedure. rhBMP-2 usage was identified via identification in the itemized billing for each procedure. Propensity scores were estimated from a multivariate logit model and matched using K Nearest Neighbors. Itemized costs were categorized using Premier Healthcare's chargemaster. Survival analysis for clinical outcomes used Kaplan-Meier Curves with log rank test and Cox Proportional Hazard models. A Bonferroni correction was established of p < 0.0035714 for 14 total clinical hypotheses tested.

RESULTS AND

CONCLUSIONS

Total initial encounter costs were 14% higher in the rhBMP-2 group, primarily driven by rhBMP-2 usage ($4,543.51). 90-day spine-related readmission rates (1.25% vs 1.39%, log-rank p = 0.447), 2-year revision fusion rates (2.7% vs 3.2%, log-rank p = 0.055), and 2-year revision decompression rates (0.29% vs 0.53%, log-rank p = 0.008) were not statistically significant after Bonferroni correction. Cox proportional hazards modeling also showed no significant protective effect of rhBMP-2 against any adverse outcome after Bonferroni correction. In addition, 2-year cumulative post-discharge costs were not significantly different between groups (p = 0.558). This may be explained by significantly higher median revision fusion costs in the rhBMP-2 group.

CONCLUSIONS

These findings suggest broad usage of rhBMP-2 may not be necessary outside of select high-risk populations. Further prospective evaluation integrating radiographic and patient reports outcomes will be essential to refining the role of rhBMP-2 in modern lumbar fusion surgery.