The use of bone morphogenetic protein in thoracolumbar spine procedures: analysis of the MarketScan longitudinal database.

BACKGROUND CONTEXT

The use of recombinant human bone morphogenetic protein (BMP) in the thoracolumbar spine remains controversial, with many questioning the risks and benefits of this new biologic.

PURPOSE

To describe national trends, incidence of complications, and revision rates associated with BMP use in thoracolumbar spine procedures.

STUDY DESIGN/SETTING: Administrative database study.

PATIENT SAMPLE

A matched cohort of 52,259 patients undergoing thoracolumbar fusion surgery from 2006 to 2010 were identified in the MarketScan database. Patients without BMP treatment were matched 2:1 to patients receiving intraoperative BMP.

OUTCOME MEASURES

Revision rates and postoperative complications.

METHODS

The MarketScan database was used to select patients undergoing thoracolumbar fusion procedures, with and without intraoperative BMP. We ascertained outcome measures using either International Classification of Disease, ninth revision, or Current Procedural Terminology coding, and matched groups were evaluated using a bivariate and multivariate analyses. Kaplan-Meier estimates of fusions failure rates were also calculated.

RESULTS

Patients receiving intraoperative BMP underwent fewer refusions, decompressions, posterior and anterior revisions, or any revision procedure (single level 4.53% vs. 5.85%, p<.0001; multilevel 5.02% vs. 6.83%, p<.0001; overall cohort 4.73% vs. 6.09%, p<.0001). After adjusting for comorbidities, demographics, and levels of procedure, BMP was not associated with the postoperative development of cancer (odds ratio 0.92). Bone morphogenetic protein use was associated with an increase in any complication at 30 days (15.8% vs. 14.9%, p=.0065), which is only statistically significant among multilevel procedures (19.74% vs. 18.02%, p=.0013). Thirty-day complications in multilevel procedures associated with BMP use included new dysrhythmia (4.68% vs. 4.01%, p=.0161) and delirium (1.08% vs. 0.69%, p=.0024). A new diagnosis of chronic pain was associated with BMP use in both single-level (2.74% vs. 2.15%, p=.0019) and multilevel (3.7% vs. 2.52%, p<.0001) procedures. Bone morphogenetic protein was negatively associated with infection in single-level procedures (2.12% vs. 2.64%, p=.0067) and wound dehiscence in multilevel procedures (0.84% vs. 1.18%, p=.0167).

CONCLUSIONS

In national data analysis of thoracolumbar procedures, we found that BMP was associated with decreased incidence of revision spinal surgery and with a slight increased risk of overall complications at 30 days. Although no BMP-associated increased risk of malignancy was found, lack of long-term follow-up precludes detection of between-group differences in malignancies and other rare events that may not appear until later.